id
float64 2.21k
2.31k
| content
stringlengths 78
1.92k
⌀ | category
stringclasses 8
values |
|---|---|---|
2,312.02419 |
Learning from human demonstrations is an emerging trend for designing intelligent robotic systems. However, previous methods typically regard videos as instructions, simply dividing them into action sequences for robotic repetition, which poses obstacles to generalization to diverse tasks or object instances. In this paper, we propose a different perspective, considering human demonstration videos not as mere instructions, but as a source of knowledge for robots. Motivated by this perspective and the remarkable comprehension and generalization capabilities exhibited by large language models (LLMs), we propose DigKnow, a method that DIstills Generalizable KNOWledge with a hierarchical structure. Specifically, DigKnow begins by converting human demonstration video frames into observation knowledge. This knowledge is then subjected to analysis to extract human action knowledge and further distilled into pattern knowledge compassing task and object instances, resulting in the acquisition of generalizable knowledge with a hierarchical structure. In settings with different tasks or object instances, DigKnow retrieves relevant knowledge for the current task and object instances. Subsequently, the LLM-based planner conducts planning based on the retrieved knowledge, and the policy executes actions in line with the plan to achieve the designated task. Utilizing the retrieved knowledge, we validate and rectify planning and execution outcomes, resulting in a substantial enhancement of the success rate. Experimental results across a range of tasks and scenes demonstrate the effectiveness of this approach in facilitating real-world robots to accomplish tasks with the knowledge derived from human demonstrations.
|
cs
|
2,312.0451 |
Recent work has shown that energy-based language modeling is an effective framework for controllable text generation because it enables flexible integration of arbitrary discriminators. However, because energy-based LMs are globally normalized, approximate techniques like Metropolis-Hastings (MH) are required for inference. Past work has largely explored simple proposal distributions that modify a single token at a time, like in Gibbs sampling. In this paper, we develop a novel MH sampler that, in contrast, proposes re-writes of the entire sequence in each step via iterative prompting of a large language model. Our new sampler (a) allows for more efficient and accurate sampling from a target distribution and (b) allows generation length to be determined through the sampling procedure rather than fixed in advance, as past work has required. We perform experiments on two controlled generation tasks, showing both downstream performance gains and more accurate target distribution sampling in comparison with single-token proposal techniques.
|
cs
|
2,312.06219 |
Predicting the future trajectories of dynamic agents in complex environments is crucial for a variety of applications, including autonomous driving, robotics, and human-computer interaction. It is a challenging task as the behavior of the agent is unknown and intrinsically multimodal. Our key insight is that the agents behaviors are influenced not only by their past trajectories and their interaction with their immediate environment but also largely with their long term waypoint (LTW). In this paper, we study the impact of adding a long-term goal on the performance of a trajectory prediction framework. We present an interpretable long term waypoint-driven prediction framework (WayDCM). WayDCM first predict an agent's intermediate goal (IG) by encoding his interactions with the environment as well as his LTW using a combination of a Discrete choice Model (DCM) and a Neural Network model (NN). Then, our model predicts the corresponding trajectories. This is in contrast to previous work which does not consider the ultimate intent of the agent to predict his trajectory. We evaluate and show the effectiveness of our approach on the Waymo Open dataset.
|
cs
|
2,312.04506 |
We show that on convex domains with sufficiently smooth boundary the limit set of non-visible Kobayashi geodesics are contained in a complex face. In two dimensions, this implies the existence of a complex tangential line segment of non-Goldilocks points in the boundary. Conversely, we construct sequences of non-visible almost-geodesics on convex domains whose boundary contains a complex tangential line segment of non-Goldilocks points in a specific direction.
|
math
|
2,312.0224 |
Improving the performance of semantic segmentation models using multispectral information is crucial, especially for environments with low-light and adverse conditions. Multi-modal fusion techniques pursue either the learning of cross-modality features to generate a fused image or engage in knowledge distillation but address multimodal and missing modality scenarios as distinct issues, which is not an optimal approach for multi-sensor models. To address this, a novel multi-modal fusion approach called CSK-Net is proposed, which uses a contrastive learning-based spectral knowledge distillation technique along with an automatic mixed feature exchange mechanism for semantic segmentation in optical (EO) and infrared (IR) images. The distillation scheme extracts detailed textures from the optical images and distills them into the optical branch of CSK-Net. The model encoder consists of shared convolution weights with separate batch norm (BN) layers for both modalities, to capture the multi-spectral information from different modalities of the same objects. A Novel Gated Spectral Unit (GSU) and mixed feature exchange strategy are proposed to increase the correlation of modality-shared information and decrease the modality-specific information during the distillation process. Comprehensive experiments show that CSK-Net surpasses state-of-the-art models in multi-modal tasks and for missing modalities when exclusively utilizing IR data for inference across three public benchmarking datasets. For missing modality scenarios, the performance increase is achieved without additional computational costs compared to the baseline segmentation models.
|
cs
|
2,312.05644 |
Designing an autonomous precise controller for ships requires accurate and reliable ship models, including the ship dynamic model and actuator model. However, selecting a suitable model for controller design and determining its parameters pose a significant challenge, considering factors such as ship actuation, input constraints, environmental disturbances, and others. This challenge is further amplified for underactuated ships, as obtaining decoupled experiment data is not feasible, and the limited data available may not adequately represent the motion characteristics of ships. To address this issue, we propose a novel model-based parameter estimation approach, called MBPE-LOGO, for underactuated ship motion models. This method combines local optimization and global optimization methods to solve the model identification problem using a dataset generated from real experiments. The effectiveness of the identified model is verified through extensive comparisons of different trajectories and prediction steps.
|
math
|
2,312.02915 |
We study the design of scheduling logic and control logic for networked control systems (NCSs) where plants communicate with their remotely located controllers over a shared band-limited communication network. Due to a limited capacity of the network, only a subset of the plants can exchange information with their controllers at any instant of time and the remaining plants operate in open-loop. Our key contribution is a new algorithm that co-designs (a) an allocation scheme of the communication network among the plants (scheduling logic) and (b) the control inputs for the plants accessing the network (control logic) under which given non-zero initial states are steered to zero in a given time horizon for all the plants in the NCS. Sparse optimization is the primary apparatus for our analysis. We also provide sufficient conditions on the plant dynamics, capacity of the communication network and the given time horizon that lead to a numerically tractable implementation of our algorithm. A numerical experiment is presented to demonstrate the proposed results.
|
eess
|
2,312.04751 |
We present the kinematic anaylsis of $246$ stars within $4^\prime$ from the center of Orion Nebula Cluster (ONC), the closest massive star cluster with active star formation across the full mass range, which provides valuable insights in the the formation and evolution of star cluster on an individual-star basis. High-precision radial velocities and surface temperatures are retrieved from spectra acquired by the NIRSPEC instrument used with adaptive optics (NIRSPAO) on the Keck II 10-m telescope. A three-dimensional kinematic map is then constructed by combining with the proper motions previously measured by the Hubble Space Telescope (HST) ACS/WFPC2/WFC3IR and Keck II NIRC2. The measured root-mean-squared velocity dispersion is $2.26\pm0.08~\mathrm{km}\,\mathrm{s}^{-1}$, significantly higher than the virial equilibrium's requirement of $1.73~\mathrm{km}\,\mathrm{s}^{-1}$, suggesting that the ONC core is supervirial, consistent with previous findings. Energy equipartition is not detected in the cluster. Most notably, the velocity of each star relative to its neighbors is found to be negatively correlated with stellar mass. Low-mass stars moving faster than their surrounding stars in a supervirial cluster suggests that the initial masses of forming stars may be related to their initial kinematic states. Additionally, a clockwise rotation preference is detected. A weak sign of inverse mass segregation is also identified among stars excluding the Trapezium stars, though it could be a sample bias. Finally, this study reports the discovery of four new candidate spectroscopic binary systems.
|
astro-ph
|
2,312.03289 |
Adversarial training integrates adversarial examples during model training to enhance robustness. However, its application in fixed dataset settings differs from real-world dynamics, where data accumulates incrementally. In this study, we investigate Adversarially Robust Class Incremental Learning (ARCIL), a method that combines adversarial robustness with incremental learning. We observe that combining incremental learning with naive adversarial training easily leads to a loss of robustness. We discover that this is attributed to the disappearance of the flatness of the loss function, a characteristic of adversarial training. To address this issue, we propose the Flatness Preserving Distillation (FPD) loss that leverages the output difference between adversarial and clean examples. Additionally, we introduce the Logit Adjustment Distillation (LAD) loss, which adapts the model's knowledge to perform well on new tasks. Experimental results demonstrate the superiority of our method over approaches that apply adversarial training to existing incremental learning methods, which provides a strong baseline for incremental learning on adversarial robustness in the future. Our method achieves AutoAttack accuracy that is 5.99\%p, 5.27\%p, and 3.90\%p higher on average than the baseline on split CIFAR-10, CIFAR-100, and Tiny ImageNet, respectively. The code will be made available.
|
cs
|
2,312.0216 |
The A-channel is a noiseless multiple access channel in which users simultaneously transmit Q-ary symbols and the receiver observes the set of transmitted symbols, but not their multiplicities. An A-channel is said to be unsourced if, additionally, users transmissions are encoded across time using a common codebook and decoding of the transmitted messages is done without regard to the identities of the active users. An interesting variant of the unsourced A-channel is the unsourced A-channel with erasures (UACE), in which transmitted symbols are erased with a given independent and identically distributed probability. In this paper, we focus on designing a code that enables a list of transmitted codewords to be recovered despite the erasures of some of the transmitted symbols. To this end, we propose the linked-loop code (LLC), which uses parity bits to link each symbol to the previous M symbols in a tail-biting manner, i.e., the first symbols of the transmission are linked to the last ones. The decoding process occurs in two phases: the first phase decodes the codewords that do not suffer from any erasures, and the second phase attempts to recover the erased symbols using the available parities. We compare the performance of the LLC over the UACE with other codes in the literature and argue for the effectiveness of the construction. Our motivation for studying the UACE comes from its relevance in machine-type communication and coded compressed sensing.
|
cs
|
2,312.05885 |
This paper focuses on parameter selection issues of kernel ridge regression (KRR). Due to special spectral properties of KRR, we find that delicate subdivision of the parameter interval shrinks the difference between two successive KRR estimates. Based on this observation, we develop an early-stopping type parameter selection strategy for KRR according to the so-called Lepskii-type principle. Theoretical verifications are presented in the framework of learning theory to show that KRR equipped with the proposed parameter selection strategy succeeds in achieving optimal learning rates and adapts to different norms, providing a new record of parameter selection for kernel methods.
|
cs
|
2,312.02896 |
Large Multimodal Models (LMMs) such as GPT-4V and LLaVA have shown remarkable capabilities in visual reasoning with common image styles. However, their robustness against diverse style shifts, crucial for practical applications, remains largely unexplored. In this paper, we propose a new benchmark, BenchLMM, to assess the robustness of LMMs against three different styles: artistic image style, imaging sensor style, and application style, where each style has five sub-styles. Utilizing BenchLMM, we comprehensively evaluate state-of-the-art LMMs and reveal: 1) LMMs generally suffer performance degradation when working with other styles; 2) An LMM performs better than another model in common style does not guarantee its superior performance in other styles; 3) LMMs' reasoning capability can be enhanced by prompting LMMs to predict the style first, based on which we propose a versatile and training-free method for improving LMMs; 4) An intelligent LMM is expected to interpret the causes of its errors when facing stylistic variations. We hope that our benchmark and analysis can shed new light on developing more intelligent and versatile LMMs.
|
cs
|
2,312.05396 |
Five distinct full-length mineralocorticoid receptor (MR) genes have been identified in humans. These human MRs can be distinguished by the presence or absence of an in-frame insertion of 12 base pairs coding for Lys, Cys, Ser, Trp (KCSW) in their DNA-binding domain (DBD) and the presence of two amino acid mutations in their amino terminal domain (NTD). Two human MRs with the KCSW insertion (MR-KCSW) and three human MRs without KCSW in the DBD have been identified. The three human MRs without KCSW contain either (Ile-180, Ala-241) or (Val-180, Val-241) or (Ile-180, Val-241) in their NTD. The two human MRs with KCSW contain either (Val-180, Val-241) or (Ile-180, Val-241) in their NTD. Human MR-KCSW with (Ile-180, Ala-241) has not been cloned. In contrast, chimpanzees contain two MRs with KCSW and two MRs without KCSW in their DBD and both contain only Ile180, Val-241 in their NTDs. Each pair of chimpanzee MRs differ at another amino acid in the NTD. A chimpanzee MR with either Val-180, Val-241 or Ile-180, Ala-241 in the NTD has not been cloned. Gorillas and orangutans each contain one MR with KCSW in the DBD and one MR without KCSW. Both gorilla and orangutan MRs contain I-180, Val-241 in their NTD. Neither Val-180, Val-241 nor Ile-180, Ala-241 are found in the NTD in either a gorilla MR or an orangutan MR. These data suggest that human MRs with Val-180, Val-241 or Ile-180, Ala-241 in the NTD evolved after humans and chimpanzees diverged from their common ancestor. These unique human MRs may have had a role in the divergent evolution of humans from chimpanzees. Studies are underway to characterize transcriptional activation of the five human MRs by aldosterone, cortisol, and other corticosteroids for comparison with each other to elucidate the roles of these MRs in human physiology.
|
q-bio
|
2,312.03337 |
We investigate generalized versions of the Iteratively Regularized Landweber Method, initially introduced in [Appl. Math. Optim., 38(1):45-68, 1998], to address linear and nonlinear ill-posed problems. Our approach is inspired by the data-driven perspective emphasized in the introduction by Aspri et al. [Numer. Funct. Anal. Optim., 41(10):1190-1227, 2020]. We provide a rigorous analysis establishing convergence and stability results and present numerical outcomes for linear operators, with the Radon transform serving as a prototype.
|
cs
|
2,312.05491 |
Captum is a comprehensive library for model explainability in PyTorch, offering a range of methods from the interpretability literature to enhance users' understanding of PyTorch models. In this paper, we introduce new features in Captum that are specifically designed to analyze the behavior of generative language models. We provide an overview of the available functionalities and example applications of their potential for understanding learned associations within generative language models.
|
cs
|
2,312.02669 |
Recent advances in metasurface lenses (metalenses) have shown great potential for opening a new era in compact imaging, photography, light detection and ranging (LiDAR), and virtual reality/augmented reality (VR/AR) applications. However, the fundamental trade-off between broadband focusing efficiency and operating bandwidth limits the performance of broadband metalenses, resulting in chromatic aberration, angular aberration, and a relatively low efficiency. In this study, a deep-learning-based image restoration framework is proposed to overcome these limitations and realize end-to-end metalens imaging, thereby achieving aberration-free full-color imaging for massproduced metalenses with 10-mm diameter. Neural network-assisted metalens imaging achieved a high resolution comparable to that of the ground truth image.
|
eess
|
2,312.0594 |
The present informal set of notes covers the material that has been presented by the author in a series of lectures for the Doctoral School in Mathematics of the Southern Federal State University of Rostov-on-Don in the Fall of 2020. The aim is to present some elementary properties of Morrey spaces and some corresponding approximation results by smooth functions.
|
math
|
2,312.05409 |
Tracking biosignals is crucial for monitoring wellness and preempting the development of severe medical conditions. Today, wearable devices can conveniently record various biosignals, creating the opportunity to monitor health status without disruption to one's daily routine. Despite widespread use of wearable devices and existing digital biomarkers, the absence of curated data with annotated medical labels hinders the development of new biomarkers to measure common health conditions. In fact, medical datasets are usually small in comparison to other domains, which is an obstacle for developing neural network models for biosignals. To address this challenge, we have employed self-supervised learning using the unlabeled sensor data collected under informed consent from the large longitudinal Apple Heart and Movement Study (AHMS) to train foundation models for two common biosignals: photoplethysmography (PPG) and electrocardiogram (ECG) recorded on Apple Watch. We curated PPG and ECG datasets from AHMS that include data from ~141K participants spanning ~3 years. Our self-supervised learning framework includes participant level positive pair selection, stochastic augmentation module and a regularized contrastive loss optimized with momentum training, and generalizes well to both PPG and ECG modalities. We show that the pre-trained foundation models readily encode information regarding participants' demographics and health conditions. To the best of our knowledge, this is the first study that builds foundation models using large-scale PPG and ECG data collected via wearable consumer devices $\unicode{x2013}$ prior works have commonly used smaller-size datasets collected in clinical and experimental settings. We believe PPG and ECG foundation models can enhance future wearable devices by reducing the reliance on labeled data and hold the potential to help the users improve their health.
|
cs
|
2,312.06015 |
In astrophysics, solving complex chemical reaction networks is essential but computationally demanding due to the high dimensionality and stiffness of the ODE systems. Traditional approaches for reducing computational load are often specialized to specific chemical networks and require expert knowledge. This paper introduces a machine learning-based solution employing autoencoders for dimensionality reduction and a latent space neural ODE solver to accelerate astrochemical reaction network computations. Additionally, we propose a cost-effective latent space linear function solver as an alternative to neural ODEs. These methods are assessed on a dataset comprising 29 chemical species and 224 reactions. Our findings demonstrate that the neural ODE achieves a 55x speedup over the baseline model while maintaining significantly higher accuracy by up to two orders of magnitude reduction in relative error. Furthermore, the linear latent model enhances accuracy and achieves a speedup of up to 4000x compared to standard methods.
|
astro-ph
|
2,312.05968 |
Purpose: General consensus amongst researchers and industry points to a lack of large, representative annotated datasets as the biggest obstacle to progress in the field of surgical data science. Self-supervised learning represents a solution to part of this problem, removing the reliance on annotations. However, the robustness of current self-supervised learning methods to domain shifts remains unclear, limiting our understanding of its utility for leveraging diverse sources of surgical data. Methods: In this work, we employ self-supervised learning to flexibly leverage diverse surgical datasets, thereby learning taskagnostic representations that can be used for various surgical downstream tasks. Based on this approach, to elucidate the impact of pre-training on downstream task performance, we explore 22 different pre-training dataset combinations by modulating three variables: source hospital, type of surgical procedure, and pre-training scale (number of videos). We then finetune the resulting model initializations on three diverse downstream tasks: namely, phase recognition and critical view of safety in laparoscopic cholecystectomy and phase recognition in laparoscopic hysterectomy. Results: Controlled experimentation highlights sizable boosts in performance across various tasks, datasets, and labeling budgets. However, this performance is intricately linked to the composition of the pre-training dataset, robustly proven through several study stages. Conclusion: The composition of pre-training datasets can severely affect the effectiveness of SSL methods for various downstream tasks and should critically inform future data collection efforts to scale the application of SSL methodologies.
Keywords: Self-Supervised Learning, Transfer Learning, Surgical Computer Vision, Endoscopic Videos, Critical View of Safety, Phase Recognition
|
cs
|
2,312.02347 |
In this paper, we give a further study in-depth of the pseudo $n$-strong Drazin inverses in an associative unital ring $R$. The characterizations of elements $a,b\in R$ for which $aa^{\tiny{\textcircled{\qihao D}}}=bb^{\tiny{\textcircled{\qihao D}}}$ are provided, and some new equivalent conditions on pseudo $n$-strong Drazin inverses are obtained. In particular, we show that an element $a\in R$ is pseudo $n$-strong Drazin invertible if, and only if, $a$ is $p$-Drazin invertible and $a-a^{n+1}\in \sqrt{J(R)}$ if, and only if, there exists $e^2=e\in {\rm comm}^2(a)$ such that $ae\in \sqrt{J(R)}$ and $1-(a+e)^n\in \sqrt{J(R)}$. We also consider pseudo $n$-strong Drazin inverses with involution, and discuss the extended versions of Cline's formula and Jacobson's lemma of this new class of generalized inverses. Likewise, we define and explore the so-called {\it pseudo $\pi$-polar} rings and demonstrate their relationships with periodic rings and strongly $\pi$-regular rings, respectively.
|
math
|
2,312.05248 |
Decentralised learning has recently gained traction as an alternative to federated learning in which both data and coordination are distributed over its users. To preserve the confidentiality of users' data, decentralised learning relies on differential privacy, multi-party computation, or a combination thereof. However, running multiple privacy-preserving summations in sequence may allow adversaries to perform reconstruction attacks. Unfortunately, current reconstruction countermeasures either cannot trivially be adapted to the distributed setting, or add excessive amounts of noise.
In this work, we first show that passive honest-but-curious adversaries can reconstruct other users' private data after several privacy-preserving summations. For example, in subgraphs with 18 users, we show that only three passive honest-but-curious adversaries succeed at reconstructing private data 11.0% of the time, requiring an average of 8.8 summations per adversary. The success rate is independent of the size of the full network. We consider weak adversaries, who do not control the graph topology and can exploit neither the workings of the summation protocol nor the specifics of users' data.
We develop a mathematical understanding of how reconstruction relates to topology and propose the first topology-based decentralised defence against reconstruction attacks. Specifically, we show that reconstruction requires a number of adversaries linear in the length of the network's shortest cycle. Consequently, reconstructing private data from privacy-preserving summations is impossible in acyclic networks.
Our work is a stepping stone for a formal theory of decentralised reconstruction defences based on topology. Such a theory would generalise our countermeasure beyond summation, define confidentiality in terms of entropy, and describe the effects of (topology-aware) differential privacy.
|
cs
|
2,312.0519 |
This paper introduces a novel approach to the fine alignment of images in a burst captured by a handheld camera. In contrast to traditional techniques that estimate two-dimensional transformations between frame pairs or rely on discrete correspondences, the proposed algorithm establishes dense correspondences by optimizing both the camera motion and surface depth and orientation at every pixel. This approach improves alignment, particularly in scenarios with parallax challenges. Extensive experiments with synthetic bursts featuring small and even tiny baselines demonstrate that it outperforms the best optical flow methods available today in this setting, without requiring any training. Beyond enhanced alignment, our method opens avenues for tasks beyond simple image restoration, such as depth estimation and 3D reconstruction, as supported by promising preliminary results. This positions our approach as a versatile tool for various burst image processing applications.
|
cs
|
2,312.02619 |
Graph contrastive learning (GCL) has emerged as a representative paradigm in graph self-supervised learning, where negative samples are commonly regarded as the key to preventing model collapse and producing distinguishable representations. Recent studies have shown that GCL without negative samples can achieve state-of-the-art performance as well as scalability improvement, with bootstrapped graph latent (BGRL) as a prominent step forward. However, BGRL relies on a complex architecture to maintain the ability to scatter representations, and the underlying mechanisms enabling the success remain largely unexplored. In this paper, we introduce an instance-level decorrelation perspective to tackle the aforementioned issue and leverage it as a springboard to reveal the potential unnecessary model complexity within BGRL. Based on our findings, we present SGCL, a simple yet effective GCL framework that utilizes the outputs from two consecutive iterations as positive pairs, eliminating the negative samples. SGCL only requires a single graph augmentation and a single graph encoder without additional parameters. Extensive experiments conducted on various graph benchmarks demonstrate that SGCL can achieve competitive performance with fewer parameters, lower time and space costs, and significant convergence speedup.
|
cs
|
2,312.04311 |
Discovering patterns in data that best describe the differences between classes allows to hypothesize and reason about class-specific mechanisms. In molecular biology, for example, this bears promise of advancing the understanding of cellular processes differing between tissues or diseases, which could lead to novel treatments. To be useful in practice, methods that tackle the problem of finding such differential patterns have to be readily interpretable by domain experts, and scalable to the extremely high-dimensional data.
In this work, we propose a novel, inherently interpretable binary neural network architecture DIFFNAPS that extracts differential patterns from data. DiffNaps is scalable to hundreds of thousands of features and robust to noise, thus overcoming the limitations of current state-of-the-art methods in large-scale applications such as in biology. We show on synthetic and real world data, including three biological applications, that, unlike its competitors, DiffNaps consistently yields accurate, succinct, and interpretable class descriptions
|
cs
|
2,312.03307 |
In the process of training a generative model, it becomes essential to measure the discrepancy between two high-dimensional probability distributions: the generative distribution and the ground-truth distribution of the observed dataset. Recently, there has been growing interest in an approach that involves slicing high-dimensional distributions, with the Cramer-Wold distance emerging as a promising method. However, we have identified that the Cramer-Wold distance primarily focuses on joint distributional learning, whereas understanding marginal distributional patterns is crucial for effective synthetic data generation. In this paper, we introduce a novel measure of dissimilarity, the mixture Cramer-Wold distance. This measure enables us to capture both marginal and joint distributional information simultaneously, as it incorporates a mixture measure with point masses on standard basis vectors. Building upon the mixture Cramer-Wold distance, we propose a new generative model called CWDAE (Cramer-Wold Distributional AutoEncoder), which shows remarkable performance in generating synthetic data when applied to real tabular datasets. Furthermore, our model offers the flexibility to adjust the level of data privacy with ease.
|
cs
|
2,312.04127 |
Extensive work has been devoted to improving the safety mechanism of Large Language Models (LLMs). However, in specific scenarios, LLMs still generate harmful responses when faced with malicious instructions, a phenomenon referred to as "Jailbreak Attack". In our research, we introduce a novel jailbreak attack method (\textbf{RADIAL}), which consists of two steps: 1) Inherent Response Tendency Analysis: we analyze the inherent affirmation and rejection tendency of LLMs to react to real-world instructions. 2) Real-World Instructions-Driven Jailbreak: based on our analysis, we strategically choose several real-world instructions and embed malicious instructions into them to amplify the LLM's potential to generate harmful responses. On three open-source human-aligned LLMs, our method achieves excellent jailbreak attack performance for both Chinese and English malicious instructions. Besides, we guided detailed ablation experiments and verified the effectiveness of our core idea "Inherent Response Tendency Analysis". Our exploration also exposes the vulnerability of LLMs to being induced into generating more detailed harmful responses in subsequent rounds of dialogue.
|
cs
|
2,312.05382 |
Differentiating noisy, discrete measurements in order to fit an ordinary differential equation can be unreasonably effective. Assuming square-integrable noise and minimal flow regularity, we construct and analyze a finite-difference differentiation filter and a Tikhonov-regularized least squares estimator for the continuous-time parameter-linear system. Combining these contributions in series, we obtain a finite-sample bound on mean absolute error of estimation. As a by-product, we offer a novel analysis of stochastically perturbed Moore-Penrose pseudoinverses.
|
cs
|
2,312.04647 |
In the paper we consider models of generalized counting processes time-changed by a general inverse subordinator, we characterize their distributions and present governing equations for them. The equations are given in terms of the generalized fractional derivatives, namely, convolutions-type derivatives with respect to Bernštein functions. Some particular examples are presented.
|
math
|
2,312.06096 |
Let $\langle A\rangle$ be the numerical semigroup generated by relatively prime positive integers $\{a_1,a_2,...,a_n\}$. The quotient of $\langle A\rangle$ with respect to a positive integer $p$ is defined by $\frac{\langle A\rangle}{p}=\{x\in \mathbb{N} \mid px\in \langle A\rangle\}$. The quotient $\frac{\langle A\rangle}{p}$ is known to be a semigroup but is hard to study. When $p$ is a positive divisor of $a_1$, we reduce the computation of the Apéry set of $\frac{a_1}{p}$ in $\frac{\langle A\rangle}{p}$ to a simple minimization problem. This allow us to obtain closed formulas of the Frobenius number of the quotient for some special numerical semigroups. These includes the cases when $\langle A\rangle$ is the almost arithmetic progressions, the almost arithmetic progressions with initial gaps, etc. In particular, we partially solve an open problem proposed by A. Adeniran et al.
|
math
|
2,312.05616 |
Real-world image super-resolution (RWSR) is a long-standing problem as low-quality (LQ) images often have complex and unidentified degradations. Existing methods such as Generative Adversarial Networks (GANs) or continuous diffusion models present their own issues including GANs being difficult to train while continuous diffusion models requiring numerous inference steps. In this paper, we propose an Iterative Token Evaluation and Refinement (ITER) framework for RWSR, which utilizes a discrete diffusion model operating in the discrete token representation space, i.e., indexes of features extracted from a VQGAN codebook pre-trained with high-quality (HQ) images. We show that ITER is easier to train than GANs and more efficient than continuous diffusion models. Specifically, we divide RWSR into two sub-tasks, i.e., distortion removal and texture generation. Distortion removal involves simple HQ token prediction with LQ images, while texture generation uses a discrete diffusion model to iteratively refine the distortion removal output with a token refinement network. In particular, we propose to include a token evaluation network in the discrete diffusion process. It learns to evaluate which tokens are good restorations and helps to improve the iterative refinement results. Moreover, the evaluation network can first check status of the distortion removal output and then adaptively select total refinement steps needed, thereby maintaining a good balance between distortion removal and texture generation. Extensive experimental results show that ITER is easy to train and performs well within just 8 iterative steps. Our codes will be available publicly.
|
cs
|
2,312.03505 |
The Event Horizon Telescope (EHT) observations carried out in 2018 April at 1.3 mm wavelengths included 9 stations in the array, comprising 7 single-dish telescopes and 2 phased arrays. The metadata package for the 2018 EHT observing campaign contains calibration tables required for the a-priori amplitude calibration of the 2018 April visibility data. This memo is the official documentation accompanying the release of the 2018 EHT metadata package, providing an overview of the contents of the package. We describe how telescope sensitivities, gain curves and other relevant parameters for each station in the EHT array were collected, processed, and validated to produce the calibration tables.
|
astro-ph
|
2,312.0655 |
The recent surge in open-source Large Language Models (LLMs), such as LLaMA, Falcon, and Mistral, provides diverse options for AI practitioners and researchers. However, most LLMs have only released partial artifacts, such as the final model weights or inference code, and technical reports increasingly limit their scope to high-level design choices and surface statistics. These choices hinder progress in the field by degrading transparency into the training of LLMs and forcing teams to rediscover many details in the training process. We present LLM360, an initiative to fully open-source LLMs, which advocates for all training code and data, model checkpoints, and intermediate results to be made available to the community. The goal of LLM360 is to support open and collaborative AI research by making the end-to-end LLM training process transparent and reproducible by everyone. As a first step of LLM360, we release two 7B parameter LLMs pre-trained from scratch, Amber and CrystalCoder, including their training code, data, intermediate checkpoints, and analyses (at ). We are committed to continually pushing the boundaries of LLMs through this open-source effort. More large-scale and stronger models are underway and will be released in the future.
|
cs
|
2,312.0347 |
We investigate the matroid realization space of a specific deformation of the regular $n$-gon with its lines of symmetry. It turns out that these particular realization spaces are birational to the elliptic modular surfaces $\Xi_{1}(n)$ over the modular curve $X_1(n)$.
We obtain in that way a model of $\Xi_{1}(n)$ defined over the rational numbers. Furthermore, a natural geometric operator acts on these matroid realizations. On the elliptic modular surface this operator corresponds to the multiplication by $-2$ on the elliptic curves. That gives a new geometric way to compute the multiplication by $-2$ on elliptic curves.
|
math
|
2,312.05337 |
What do artificial neural networks (ANNs) learn? The machine learning (ML) community shares the narrative that ANNs must develop abstract human concepts to perform complex tasks. Some go even further and believe that these concepts are stored in individual units of the network. Based on current research, I systematically investigate the assumptions underlying this narrative. I conclude that ANNs are indeed capable of performing complex prediction tasks, and that they may learn human and non-human concepts to do so. However, evidence indicates that ANNs do not represent these concepts in individual units.
|
cs
|
2,312.03767 |
Open Set Domain Adaptation (OSDA) aims to adapt a model trained on a source domain to a target domain that undergoes distribution shift and contains samples from novel classes outside the source domain. Source-free OSDA (SF-OSDA) techniques eliminate the need to access source domain samples, but current SF-OSDA methods utilize only the known classes in the target domain for adaptation, and require access to the entire target domain even during inference after adaptation, to make the distinction between known and unknown samples. In this paper, we introduce Unknown Sample Discovery (USD) as an SF-OSDA method that utilizes a temporally ensembled teacher model to conduct known-unknown target sample separation and adapts the student model to the target domain over all classes using co-training and temporal consistency between the teacher and the student. USD promotes Jensen-Shannon distance (JSD) as an effective measure for known-unknown sample separation. Our teacher-student framework significantly reduces error accumulation resulting from imperfect known-unknown sample separation, while curriculum guidance helps to reliably learn the distinction between target known and target unknown subspaces. USD appends the target model with an unknown class node, thus readily classifying a target sample into any of the known or unknown classes in subsequent post-adaptation inference stages. Empirical results show that USD is superior to existing SF-OSDA methods and is competitive with current OSDA models that utilize both source and target domains during adaptation.
|
cs
|
2,312.04969 |
An accurate delay and Doppler estimation method for a radar system using time and frequency-shifted pulses with pseudo-random numbers is proposed. The ambiguity function of the transmitted signal has a strong peak at the origin and is close to zero if delay and Doppler are more than the inverses of the bandwidth and time-width. A two-dimensional (2D) sinc function gives a good approximation of the ambiguity function around the origin, by which fractional delay and Doppler are accurately estimated.
|
cs
|
2,312.02845 |
Since the demise of the password was predicted in 2004, different attempts in industry and academia have been made to create an alternative for the use of passwords in authentication, without compromising on security and user experience. This review examines password-less authentication schemes that have been proposed since after the death knell was placed on passwords in 2004. We start with a brief discussion of the requirements of authentication systems and then identify various password-less authentication proposals to date. We then evaluate the truly password-less and practical schemes using a framework that examines authentication credentials based on their impact on user experience, overall security, and ease of deployment. The findings of this review observe a difficulty in balancing security with a user experience compared to that of passwords in new password-less schemes, providing the opportunity for new applied research to leverage existing knowledge and combine technologies and techniques in innovative ways that can address this imbalance.
|
cs
|
2,312.03522 |
We establish local Hölder estimates for viscosity solutions of fully nonlinear second order equations with quadratic growth in the gradient and unbounded right-hand side in $L^q$ spaces, for an integrability threshold $q$ guaranteeing the validity of the maximum principle. This is done through a nonlinear Harnack inequality for nonhomogeneous equations driven by a uniformly elliptic Isaacs operator and perturbed by a Hamiltonian term with natural growth in the gradient. As a byproduct, we derive a new Liouville property for entire $L^p$ viscosity solutions of fully nonlinear equations as well as a nonlinear Calderón-Zygmund estimate for strong solutions of such equations.
|
math
|
2,312.05126 |
The splashback radius, coinciding with the minimum in the dark matter radial density gradient, is thought to be a universal definition of the edge of a dark matter halo. Observational methods to detect it have traced the dark matter using weak gravitational lensing or galaxy number counts. Recent attempts have also claimed the detection of a similar feature in Sunyaev-Zel'dovich (SZ) observations of the hot intracluster gas. Here, we use the FLAMINGO simulations to investigate whether a reflection of the splashback feature is predicted to occur in the cluster gas profiles. We find that the minimum in the gradient of the stacked 3D gas density and pressure profiles, and the maximum in the entropy profile, broadly align with the splashback feature though there are significant differences. While the dark matter splashback radius varies with specific mass accretion rate, in agreement with previous work, the radial position of the feature in the gas density is more sensitive to halo mass. In addition, we show that the feature is also present in projected 2D pseudo-observable profiles: emission measure (X-ray); Compton-$y$ (SZ) and surface mass density (weak lensing). We find that the latter traces the dark matter results reasonably well albeit the feature occurs at a slightly smaller radius. While results for the gas profiles are largely insensitive to accretion rate and various observable proxies for dynamical state, they do depend on the strength of the feedback processes.
|
astro-ph
|
2,312.04131 |
The scarcity of labeled audio-visual datasets is a constraint for training superior audio-visual speaker diarization systems. To improve the performance of audio-visual speaker diarization, we leverage pre-trained supervised and self-supervised speech models for audio-visual speaker diarization. Specifically, we adopt supervised~(ResNet and ECAPA-TDNN) and self-supervised pre-trained models~(WavLM and HuBERT) as the speaker and audio embedding extractors in an end-to-end audio-visual speaker diarization~(AVSD) system. Then we explore the effectiveness of different frameworks, including Transformer, Conformer, and cross-attention mechanism, in the audio-visual decoder. To mitigate the degradation of performance caused by separate training, we jointly train the audio encoder, speaker encoder, and audio-visual decoder in the AVSD system. Experiments on the MISP dataset demonstrate that the proposed method achieves superior performance and obtained third place in MISP Challenge 2022.
|
eess
|
2,312.0268 |
Let $G$ be a graph. The numbers of positive, negative and zero eigenvalues (including multiplicities) of the adjacency matrix $A(G)$ is denoted by $p(G)$, $n(G)$ and $\eta(G)$, respectively. In general, $n(G)$ (resp. $p(G)$) is called the negative (resp. positive) inertia index of $G$. Suppose that a connected graph $G$ has at least one cycle and let $g$ be the length of the shortest cycle in $G$. In this paper, we prove $n(G)\geq\lceil \frac{g}{2}\rceil-1$. Further, the extremal graphs corresponding to $n(G)=\lceil \frac{g}{2}\rceil-1$ and $n(G)=\lceil \frac{g}{2}\rceil$ are characterized, respectively.
|
math
|
2,312.06455 |
Toman et al. have proposed a type system for automatic verification of low-level programs, which combines ownership types and refinement types to enable strong updates of refinement types in the presence of pointer aliases. We extend their type system to support pointer arithmetic, and prove its soundness. Based on the proposed type system, we have implemented a prototype tool for automated verification of the lack of assertion errors of low-level programs with pointer arithmetic, and confirmed its effectiveness through experiments.
|
cs
|
2,312.06495 |
Security is an important topic in our contemporary world, and the ability to automate the detection of any events of interest that can take place in a crowd is of great interest to a population. We hypothesize that the detection of events in videos is correlated with significant changes in pedestrian behaviors. In this paper, we examine three different scenarios of crowd behavior, containing both the cases where an event triggers a change in the behavior of the crowd and two video sequences where the crowd and its motion remain mostly unchanged. With both the videos and the tracking of the individual pedestrians (performed in a pre-processed phase), we use Geomind, a software we developed to extract significant data about the scene, in particular, the geometrical features, personalities, and emotions of each person. We then examine the output, seeking a significant change in the way each person acts as a function of the time, that could be used as a basis to identify events or to model realistic crowd actions. When applied to the games area, our method can use the detected events to find some sort of pattern to be then used in agent simulation. Results indicate that our hypothesis seems valid in the sense that the visually observed events could be automatically detected using GeoMind.
|
cs
|
2,312.06579 |
Amazon Locker is a self-service delivery or pickup location where customers can pick up packages and drop off returns. A basic first-come-first-served policy for accepting package delivery requests to lockers results in lockers becoming full with standard shipping speed (3-5 day shipping) packages, and leaving no space left for expedited packages which are mostly Next-Day or Two-Day shipping. This paper proposes a solution to the problem of determining how much locker capacity to reserve for different ship-option packages. Yield management is a much researched field with popular applications in the airline, car rental, and hotel industries. However, Amazon Locker poses a unique challenge in this field since the number of days a package will wait in a locker (package dwell time) is, in general, unknown. The proposed solution combines machine learning techniques to predict locker demand and package dwell time, and linear programming to maximize throughput in lockers. The decision variables from this optimization provide optimal capacity reservation values for different ship options. This resulted in a year-over-year increase of 9% in Locker throughput worldwide during holiday season of 2018, impacting millions of customers.
|
math
|
2,312.04241 |
This work is concerned with an inverse scattering problem of determining unknown scatterers from time-dependent acoustic measurements. A novel time-domain direct sampling method is developed to efficiently determine both the locations and shapes of inhomogeneous media. In particular, our approach is very easy to implement since only cheap space-time integrations are involved in the evaluation of the imaging functionals. Based on the Fourier-Laplace transform, we establish an inherent connection between the time-domain and frequency-domain direct sampling method. Moreover, rigorous theoretical justifications and numerical experiments are provided to verify the validity and feasibility of the proposed method.
|
math
|
2,312.0318 |
Many imaging science tasks can be modeled as a discrete linear inverse problem. Solving linear inverse problems is often challenging, with ill-conditioned operators and potentially non-unique solutions. Embedding prior knowledge, such as smoothness, into the solution can overcome these challenges. In this work, we encode prior knowledge using a non-negative patch dictionary, which effectively learns a basis from a training set of natural images. In this dictionary basis, we desire solutions that are non-negative and sparse (i.e., contain many zero entries). With these constraints, standard methods for solving discrete linear inverse problems are not directly applicable. One such approach is the modified residual norm steepest descent (MRNSD), which produces non-negative solutions but does not induce sparsity. In this paper, we provide two methods based on MRNSD that promote sparsity. In our first method, we add an $\ell_1$-regularization term with a new, optimal step size. In our second method, we propose a new non-negative, sparsity-promoting mapping of the solution. We compare the performance of our proposed methods on a number of numerical experiments, including deblurring, image completion, computer tomography, and superresolution. Our results show that these methods effectively solve discrete linear inverse problems with non-negativity and sparsity constraints.
|
math
|
2,312.03647 |
Double staining in histopathology, particularly for metaplastic breast cancer, typically employs H&E and P63 dyes. However, P63's tissue damage and high cost necessitate alternative methods. This study introduces xAI-CycleGAN, an advanced architecture combining Mask CycleGAN with explainability features and structure-preserving capabilities for transforming H&E stained breast tissue images into P63-like images. The architecture allows for output editing, enhancing resemblance to actual images and enabling further model refinement. We showcase xAI-CycleGAN's efficacy in maintaining structural integrity and generating high-quality images. Additionally, a histopathologist survey indicates the generated images' realism is often comparable to actual images, validating our model's high-quality output.
|
eess
|
2,312.0454 |
Modeling spatial-temporal interactions among neighboring agents is at the heart of multi-agent problems such as motion forecasting and crowd navigation. Despite notable progress, it remains unclear to which extent modern representations can capture the causal relationships behind agent interactions. In this work, we take an in-depth look at the causal awareness of these representations, from computational formalism to real-world practice. First, we cast doubt on the notion of non-causal robustness studied in the recent CausalAgents benchmark. We show that recent representations are already partially resilient to perturbations of non-causal agents, and yet modeling indirect causal effects involving mediator agents remains challenging. To address this challenge, we introduce a metric learning approach that regularizes latent representations with causal annotations. Our controlled experiments show that this approach not only leads to higher degrees of causal awareness but also yields stronger out-of-distribution robustness. To further operationalize it in practice, we propose a sim-to-real causal transfer method via cross-domain multi-task learning. Experiments on pedestrian datasets show that our method can substantially boost generalization, even in the absence of real-world causal annotations. We hope our work provides a new perspective on the challenges and potential pathways towards causally-aware representations of multi-agent interactions. Our code is available at .
|
cs
|
2,312.06268 |
As the complexity of digital circuits increases, High-Level Synthesis (HLS) is becoming a valuable tool to increase productivity and design reuse by utilizing relevant Electronic Design Automation (EDA) flows, either for Application-Specific Integrated Circuits (ASIC) or for Field Programmable Gate Arrays (FPGA). Side Channel Analysis (SCA) and Fault Injection (FI) attacks are powerful hardware attacks, capable of greatly weakening the theoretical security levels of secure implementations. Furthermore, critical applications demand high levels of reliability including fault tolerance. The lack of security and reliability driven optimizations in HLS tools makes it necessary for the HLS-based designs to validate that the properties of the algorithm and the countermeasures have not been compromised due to the HLS flow. In this work, we provide results on the resilience evaluation of HLS-based FPGA implementations for the aforementioned threats. As a test case, we use multiple versions of an on-the-fly SBOX algorithm integrating different countermeasures (hiding and masking), written in C and implemented using Vivado HLS. We perform extensive evaluations for all the designs and their optimization scenarios. The results provide evidence of issues arising due to HLS optimizations on the security and the reliability of cryptographic implementations. Furthermore, the results put HLS algorithms to the test of designing secure accelerators and can lead to improving them towards the goal of increasing productivity in the domain of secure and reliable cryptographic implementations.
|
cs
|
2,312.04284 |
Quantum Darwinism (QD) proposes that classical objectivity emerges from the broadcast of information about a microscopic degree of freedom into multiple fractions of a many-body environment. Such a broadcast of information is in sharp contrast with its scrambling under strong interaction. It was recently shown that quantum dynamics interpolating between broadcasting and scrambling may display sharp phase transitions of information propagation, named QD-encoding transitions. Here, we initiate their systematic study in generic, non-Clifford settings. First, in a general theoretical setup where the information propagation is modeled as an isometry, whose input qudit is entangled with a reference, we propose a probe of the transitions -- the distribution of the density matrix of the reference after measuring an environment fraction. This probe measures the classical correlation between the fraction and the injected information. We then apply the framework to two similar models defined by a tensor network on an expanding tree, modeling a noisy apparatus that attempts to broadcast the $z$ component of a spin-half. We derive an exact recursion relation of the density matrix distribution, which we analyze analytically and numerically. As a result we find three phases: QD, intermediate and encoding, and two continuous transitions. The encoding-intermediate transition describes the establishment of nonzero correlation between the reference and a small environment fraction, and can be probed by a ``coarse-grained'' measure of the total spin-$z$ of the fraction, which becomes non-Gaussian and symmetry breaking in the intermediate space. The QD-intermediate transition is about whether the correlation is perfect. It must be probed by fined-grained measures, and corresponds to a more subtle symmetry breaking in the replica space.
|
quant-ph
|
2,312.04702 |
The canonical modular homomorphism provides an embedding of $\mathbb{R}$ into the outer automorphism group Out($M$) of any type III$_{1}$ factor $M$. We provide an explicit construction of a full factor of type III$_{1}$ with separable predual such that the outer automorphism group is minimal, i.e. this embedding is an isomorphism. We obtain such a III$_{1}$ factor by using an amalgamated free product construction.
|
math
|
2,312.04571 |
Swarm-Merging, SwarMer, is a decentralized framework to localize Flying Light Specks (FLSs) to render 2D and 3D shapes. An FLS is a miniature sized drone equipped with one or more light sources to generate different colors and textures with adjustable brightness. It is battery powered, network enabled with storage and processing capability to implement a decentralized algorithm such as SwarMer. An FLS is unable to render a shape by itself. SwarMer uses the inter-FLS relationship effect of its organizational framework to compensate for the simplicity of each individual FLS, enabling a swarm of cooperating FLSs to render complex shapes. SwarMer is resilient to both FLSs failing and FLSs leaving to charge their battery. It is fast, highly accurate, and scales to remain effective when a shape consists of a large number of FLSs.
|
cs
|
2,312.0412 |
In 1950, P. Erdős proved that there is an infinite arithmetic progression of positive odd integers none of which can be represented as the sum of a prime and a power of two. Erdős conjectured that the set of positive odd integers that cannot be represented as the sum of a prime and a power of two is the union of an infinite arithmetic progression of positive odd integers and a set of asymptotic density zero. Erdős called this conjecture ``rather silly". This is Problem 16 of Bloom's list of Erdős' problems. In this paper, we prove that this conjecture is false.
|
math
|
2,312.04216 |
Despite the impressive feats demonstrated by Reinforcement Learning (RL), these algorithms have seen little adoption in high-risk, real-world applications due to current difficulties in explaining RL agent actions and building user trust. We present Counterfactual Demonstrations for Explanation (CODEX), a method that incorporates semantic clustering, which can effectively summarize RL agent behavior in the state-action space. Experimentation on the MiniGrid and StarCraft II gaming environments reveals the semantic clusters retain temporal as well as entity information, which is reflected in the constructed summary of agent behavior. Furthermore, clustering the discrete+continuous game-state latent representations identifies the most crucial episodic events, demonstrating a relationship between the latent and semantic spaces. This work contributes to the growing body of work that strives to unlock the power of RL for widespread use by leveraging and extending techniques from Natural Language Processing.
|
cs
|
2,312.05623 |
Prior works have analyzed the performance of millimeter wave automotive radars in the presence of diverse clutter and interference scenarios using stochastic geometry tools instead of more time-consuming measurement studies or system-level simulations. In these works, the distributions of radars or discrete clutter scatterers were modeled as Poisson point processes in the Euclidean space. However, since most automotive radars are likely to be mounted on vehicles and road infrastructure, road geometries are an important factor that must be considered. Instead of considering each road geometry as an individual case for study, in this work, we model each case as a specific instance of an underlying Poisson line process and further model the distribution of vehicles on the road as a Poisson point process - forming a Poisson line Cox process. Then, through the use of stochastic geometry tools, we estimate the average number of interfering radars for specific road and vehicular densities and the effect of radar parameters such as noise and beamwidth on the radar detection metrics. The numerical results are validated with Monte Carlo simulations.
|
eess
|
2,312.05146 |
We prove a quantitative version of the Gaussian Faber-Krahn type inequality proved by Betta, Chiacchio and Ferone for the first Dirichlet eigenvalue of the Ornstein-Uhlenbeck operator, estimating the deficit in terms of the Gaussian Fraenkel asymmetry. As expected, the multiplicative constant only depends on the prescribed Gaussian measure.
|
math
|
2,312.04598 |
Cooperative robots can significantly assist people in their productive activities, improving the quality of their works. Collision detection is vital to ensure the safe and stable operation of cooperative robots in productive activities. As an advanced geometric language, conformal geometric algebra can simplify the construction of the robot collision model and the calculation of collision distance. Compared with the formal method based on conformal geometric algebra, the traditional method may have some defects which are difficult to find in the modelling and calculation. We use the formal method based on conformal geometric algebra to study the collision detection problem of cooperative robots. This paper builds formal models of geometric primitives and the robot body based on the conformal geometric algebra library in HOL Light. We analyse the shortest distance between geometric primitives and prove their collision determination conditions. Based on the above contents, we construct a formal verification framework for the robot collision detection method. By the end of this paper, we apply the proposed framework to collision detection between two single-arm industrial cooperative robots. The flexibility and reliability of the proposed framework are verified by constructing a general collision model and a special collision model for two single-arm industrial cooperative robots.
|
cs
|
2,312.0378 |
Construction waste hauling trucks (CWHTs), as one of the most commonly seen heavy-duty vehicles in major cities around the globe, are usually subject to a series of regulations and spatial-temporal access restrictions because they not only produce significant NOx and PM emissions but also causes on-road fugitive dust. The timely and accurate prediction of CWHTs' destinations and dwell times play a key role in effective environmental management. To address this challenge, we propose a prediction method based on an interpretable activity-based model, input-output hidden Markov model (IOHMM), and validate it on 300 CWHTs in Chengdu, China. Contextual factors are considered in the model to improve its prediction power. Results show that the IOHMM outperforms several baseline models, including Markov chains, linear regression, and long short-term memory. Factors influencing the predictability of CWHTs' transportation activities are also explored using linear regression models. Results suggest the proposed model holds promise in assisting authorities by predicting the upcoming transportation activities of CWHTs and administering intervention in a timely and effective manner.
|
cs
|
2,312.06636 |
This paper is the third part of the series "Spherical higher order Fourier analysis over finite fields", aiming to develop the higher order Fourier analysis method along spheres over finite fields, and to solve the geometric Ramsey conjecture in the finite field setting.
In this paper, we prove an inverse theorem over finite field for spherical Gowers norms, i.e. a local Gowers norm supported on a sphere. We show that if the $(s+1)$-th spherical Gowers norm of a 1-bounded function $f\colon\mathbb{F}_{p}^{d}\to \mathbb{C}$ is at least $\epsilon$ and if $d$ is sufficiently large depending only on $s$, then $f$ correlates on the sphere with a $p$-periodic $s$-step nilsequence, where the bounds for the complexity and correlation depend only on $d$ and $\epsilon$. This result will be used in later parts of the series to prove the geometric Ramsey conjecture in the finite field setting.
|
math
|
2,312.04227 |
In the mentor-student communication process, students often struggle to receive prompt and clear guidance from their mentors, making it challenging to determine their next steps. When mentors don't respond promptly, it can lead to student confusion, as they may be uncertain whether their message has been acknowledged without resulting action. Instead of the binary options of "read" and "unread," there's a pressing need for more nuanced descriptions of message states. To tackle this ambiguity, we've developed a set of symbols to precisely represent the cognitive states associated with messages in transit. Through experimentation, this design not only assists mentors and students in effectively labeling their responses but also mitigates unnecessary misunderstandings. By utilizing symbols for accurate information and understanding state marking, we've enhanced communication efficiency between mentors and students, thereby improving the quality and efficacy of communication in mentor-student relationships.
|
cs
|
2,312.03609 |
Resilience is emerging as an evolving notion, reflecting a system's ability to endure and adapt to sudden and catastrophic changes and disruptions. This paper spotlights the significance of the quantitative resilience indices of medium-voltage DC (MVDC) distribution technology in marine vessels, notably naval ships. Given the intricate electrical requirements of modern naval ships, the need for a robust power supply underlines the imperative of resilient DC microgrids. Addressing this, our study introduces a novel quantitative metric for operational resilience of DC microgrids based on the measured voltage of main DC bus. This metric not only fuses real-time tracking, compatibility, and computational efficiency, but also adeptly monitors multiple event phases based on time-domain analysis of dc bus voltage dynamics. The intricacies of the dc bus voltage, including overshoots and undershoots, are meticulously accounted for in the algorithm design. With respect to existing research that typically focuses on offline resilience assessments, the proposed index provides valuable real-time information for microgrid operators and identifies whether microgrid resilience is deteriorating over time.
|
cs
|
2,312.0543 |
Talking face generation has gained immense popularity in the computer vision community, with various applications including AR/VR, teleconferencing, digital assistants, and avatars. Traditional methods are mainly audio-driven ones which have to deal with the inevitable resource-intensive nature of audio storage and processing. To address such a challenge, we propose FT2TF - First-Person Statement Text-To-Talking Face Generation, a novel one-stage end-to-end pipeline for talking face generation driven by first-person statement text. Moreover, FT2TF implements accurate manipulation of the facial expressions by altering the corresponding input text. Different from previous work, our model only leverages visual and textual information without any other sources (e.g. audio/landmark/pose) during inference. Extensive experiments are conducted on LRS2 and LRS3 datasets, and results on multi-dimensional evaluation metrics are reported. Both quantitative and qualitative results showcase that FT2TF outperforms existing relevant methods and reaches the state-of-the-art. This achievement highlights our model capability to bridge first-person statements and dynamic face generation, providing insightful guidance for future work.
|
cs
|
2,312.03838 |
The Hayden-Preskill protocol probes the capability of information recovery from local subsystems after unitary dynamics. As such it resolves the capability of quantum many-body systems to dynamically implement a quantum error-correcting code. The transition to coding behavior has been mostly discussed using effective approaches, such as entanglement membrane theory. Here, we present exact results on the use of Hayden-Preskill recovery as a dynamical probe of scrambling in local quantum many-body systems. We investigate certain classes of unitary circuit models, both structured Floquet (dual-unitary) and Haar-random circuits. We discuss different dynamical signatures corresponding to information transport or scrambling, respectively, that go beyond effective approaches. Surprisingly, certain chaotic circuits transport information with perfect fidelity. In integrable dual-unitary circuits, we relate the information transmission to the propagation and scattering of quasiparticles. Using numerical and analytical insights, we argue that the qualitative features of information recovery extend away from these solvable points. Our results suggest that information recovery protocols can serve to distinguish chaotic and integrable behavior, and that they are sensitive to characteristic dynamical features, such as long-lived quasiparticles or dual-unitarity.
|
quant-ph
|
2,312.022 |
Major advancements in computer vision can primarily be attributed to the use of labeled datasets. However, acquiring labels for datasets often results in errors which can harm model performance. Recent works have proposed methods to automatically identify mislabeled images, but developing strategies to effectively implement them in real world datasets has been sparsely explored. Towards improved data-centric methods for cleaning real world vision datasets, we first conduct more than 200 experiments carefully benchmarking recently developed automated mislabel detection methods on multiple datasets under a variety of synthetic and real noise settings with varying noise levels. We compare these methods to a Simple and Efficient Mislabel Detector (SEMD) that we craft, and find that SEMD performs similarly to or outperforms prior mislabel detection approaches. We then apply SEMD to multiple real world computer vision datasets and test how dataset size, mislabel removal strategy, and mislabel removal amount further affect model performance after retraining on the cleaned data. With careful design of the approach, we find that mislabel removal leads per-class performance improvements of up to 8% of a retrained classifier in smaller data regimes.
|
stat
|
2,312.03392 |
We present an analog of O'Neill's Theorem (Theorem 5.2 in [17]) for finite games, which reveals a picture of the structure of equilibria under payoff perturbations in finite games.
|
cs
|
2,312.04086 |
Recently, video generation has attracted massive attention and yielded noticeable outcomes. Concerning the characteristics of video, multi-text conditioning incorporating sequential events is necessary for next-step video generation. In this work, we propose a novel multi-text video generation~(MTVG) by directly utilizing a pre-trained diffusion-based text-to-video~(T2V) generation model without additional fine-tuning. To generate consecutive video segments, visual consistency generated by distinct prompts is necessary with diverse variations, such as motion and content-related transitions. Our proposed MTVG includes Dynamic Noise and Last Frame Aware Inversion which reinitialize the noise latent to preserve visual coherence between videos of different prompts and prevent repetitive motion or contents. Furthermore, we present Structure Guiding Sampling to maintain the global appearance across the frames in a single video clip, where we leverage iterative latent updates across the preceding frame. Additionally, our Prompt Generator allows for arbitrary format of text conditions consisting of diverse events. As a result, our extensive experiments, including diverse transitions of descriptions, demonstrate that our proposed methods show superior generated outputs in terms of semantically coherent and temporally seamless video.Video examples are available in our project page: .
|
cs
|
2,312.04745 |
In fairness audits, a standard objective is to detect whether a given algorithm performs substantially differently between subgroups. Properly powering the statistical analysis of such audits is crucial for obtaining informative fairness assessments, as it ensures a high probability of detecting unfairness when it exists. However, limited guidance is available on the amount of data necessary for a fairness audit, lacking directly applicable results concerning commonly used fairness metrics. Additionally, the consideration of unequal subgroup sample sizes is also missing. In this tutorial, we address these issues by providing guidance on how to determine the required subgroup sample sizes to maximize the statistical power of hypothesis tests for detecting unfairness. Our findings are applicable to audits of binary classification models and multiple fairness metrics derived as summaries of the confusion matrix. Furthermore, we discuss other aspects of audit study designs that can increase the reliability of audit results.
|
stat
|
2,312.03108 |
Having previously been the subject of decades of semiconductor research, cadmium arsenide has now reemerged as a topological material, realizing ideal three-dimensional Dirac points at the Fermi level. These topological Dirac points lead to a number of extraordinary transport phenomena, including strong quantum oscillations, large magnetoresistance, ultrahigh mobilities, and Fermi velocities exceeding graphene. The large mobilities persist even in thin films and nanowires of cadmium arsenide, suggesting the involvement of topological surface states. However, computational studies of the surface states in this material are lacking, in part due to the large 80-atom unit cell. Here we present the computed Fermi arc surface states of a cadmium arsenide thin film, based on a tight-binding model derived directly from the electronic structure. We show that despite the close proximity of the Dirac points, the Fermi arcs are very long and straight, extending through nearly the entire Brillouin zone. The shape and spin properties of the Fermi arcs suppress both back- and side- scattering at the surface, which we show by explicit integrals over the phase space. The introduction of a small symmetry-breaking term, expected in a strong electric field, gaps the electronic structure, creating a weak topological insulator phase that exhibits similar transport properties. Crucially, the mechanisms suppressing scattering in this material differ from those in other topological materials such as Weyl semimetals and topological insulators, suggesting a new route for engineering high-mobility devices based on Dirac semimetal surface states.
|
quant-ph
|
2,312.02349 |
The minimum orbital separation of planets in long-stable planetary systems is often modeled as a step function, parameterized with a single value $\Delta_{min}$ (measured in mutual Hill radius of the two neighboring planets). Systems with smaller separations are considered unstable, and planet pairs with greater separations are considered stable. Here we report that a log-normal distribution function for $\Delta_{min}$, rather than a single threshold value, provides a more accurate model. From our suite of simulated planetary systems, the parameters of the best-fit log-normal distribution are $\mu=1.97\pm0.02$ and $\sigma=0.40\pm0.02$, such that the mean, median, and mode of $\Delta_{min}$ are 7.77, 7.17, and 6.11, respectively. This result is consistent with previous estimates for $\Delta_{min}$ threshold values in the range $\sim$5$-$8. We find a modest dependence of the distribution of $\Delta_{min}$ on multiplicity within the system, as well as on planetary mass ratios of the closest planet pair. The overall distribution of nearest-neighbor planetary orbital spacings (measured in the mutual Hill radii and denoted simply as $\Delta$) in long-term stable systems is also well fit with a log-normal distribution, with parameters $\mu=3.14\pm0.03$ and $\sigma=0.76\pm0.02$. In simulations of sets of many planets initially packed very close together, we find that the orbital spacings of long-term stable systems is statistically similar to that in the observed Kepler sample of exo-planetary systems, indicating a strong role of sculpting of planetary architectures by dynamical instabilities.
|
astro-ph
|
2,312.04489 |
A link between first-order ordinary differential equations (ODEs) and 2-dimensional Riemannian manifolds is explored. Given a first-order ODE, an associated Riemannian metric on the variable space is defined, and some properties of the resulting surface are studied, including a connection between Jacobi fields and Lie point symmetries. In particular, it is proven that if the associated surface is flat, then the ODE can be integrated by quadratures.
Next, deformations of the associated surfaces are considered. A relation between some Jacobi fields on the deformed surface and integrability of the ODE is established, showing that there is a class of vector fields, beyond Lie point symmetries, which are useful for solving first-order ODEs. As a result, it is concluded that the deformation into a constant curvature surface leads to the integrability of the given ODE.
|
math
|
2,312.05557 |
This is the first treatise on multi-user (MU) beamforming designed for achieving long-term rate-fairness in fulldimensional MU massive multi-input multi-output (m-MIMO) systems. Explicitly, based on the channel covariances, which can be assumed to be known beforehand, we address this problem by optimizing the following objective functions: the users' signal-toleakage-noise ratios (SLNRs) using SLNR max-min optimization, geometric mean of SLNRs (GM-SLNR) based optimization, and SLNR soft max-min optimization. We develop a convex-solver based algorithm, which invokes a convex subproblem of cubic time-complexity at each iteration for solving the SLNR maxmin problem. We then develop closed-form expression based algorithms of scalable complexity for the solution of the GMSLNR and of the SLNR soft max-min problem. The simulations provided confirm the users' improved-fairness ergodic rate distributions.
|
cs
|
2,312.05794 |
Performance of ordinary least squares(OLS) method for the \emph{estimation of high dimensional stable state transition matrix} $A$(i.e., spectral radius $\rho(A)<1$) from a single noisy observed trajectory of the linear time invariant(LTI)\footnote{Linear Gaussian (LG) in Markov chain literature} system $X_{-}:(x_0,x_1, \ldots,x_{N-1})$ satisfying \begin{equation}
x_{t+1}=Ax_{t}+w_{t}, \hspace{10pt} \text{ where } w_{t} \thicksim N(0,I_{n}), \end{equation}
heavily rely on negative moments of the sample covariance matrix: $(X_{-}X_{-}^{*})=\sum_{i=0}^{N-1}x_{i}x_{i}^{*}$ and singular values of $EX_{-}^{*}$, where $E$ is a rectangular Gaussian ensemble $E=[w_0, \ldots, w_{N-1}]$. Negative moments requires sharp estimates on all the eigenvalues $\lambda_{1}\big(X_{-}X_{-}^{*}\big) \geq \ldots \geq \lambda_{n}\big(X_{-}X_{-}^{*}\big) \geq 0$. Leveraging upon recent results on spectral theorem for non-Hermitian operators in \cite{naeem2023spectral}, along with concentration of measure phenomenon and perturbation theory(Gershgorins' and Cauchys' interlacing theorem) we show that only when $A=A^{*}$, typical order of $\lambda_{j}\big(X_{-}X_{-}^{*}\big) \in \big[N-n\sqrt{N}, N+n\sqrt{N}\big]$ for all $j \in [n]$. However, in \emph{high dimensions} when $A$ has only one distinct eigenvalue $\lambda$ with geometric multiplicity of one, then as soon as eigenvalue leaves \emph{complex half unit disc}, largest eigenvalue suffers from curse of dimensionality: $\lambda_{1}\big(X_{-}X_{-}^{*}\big)=\Omega\big( \lfloor\frac{N}{n}\rfloor e^{\alpha_{\lambda}n} \big)$, while smallest eigenvalue $\lambda_{n}\big(X_{-}X_{-}^{*}\big) \in (0, N+\sqrt{N}]$. Consequently, OLS estimator incurs a \emph{phase transition} and becomes \emph{transient: increasing iteration only worsens estimation error}, all of this happening when the dynamics are generated from stable systems.
|
math
|
2,312.03951 |
The phenomenon of model-wise double descent, where the test error peaks and then reduces as the model size increases, is an interesting topic that has attracted the attention of researchers due to the striking observed gap between theory and practice \citep{Belkin2018ReconcilingMM}. Additionally, while double descent has been observed in various tasks and architectures, the peak of double descent can sometimes be noticeably absent or diminished, even without explicit regularization, such as weight decay and early stopping. In this paper, we investigate this intriguing phenomenon from the optimization perspective and propose a simple optimization-based explanation for why double descent sometimes occurs weakly or not at all. To the best of our knowledge, we are the first to demonstrate that many disparate factors contributing to model-wise double descent (initialization, normalization, batch size, learning rate, optimization algorithm) are unified from the viewpoint of optimization: model-wise double descent is observed if and only if the optimizer can find a sufficiently low-loss minimum. These factors directly affect the condition number of the optimization problem or the optimizer and thus affect the final minimum found by the optimizer, reducing or increasing the height of the double descent peak. We conduct a series of controlled experiments on random feature models and two-layer neural networks under various optimization settings, demonstrating this optimization-based unified view. Our results suggest the following implication: Double descent is unlikely to be a problem for real-world machine learning setups. Additionally, our results help explain the gap between weak double descent peaks in practice and strong peaks observable in carefully designed setups.
|
cs
|
2,312.03019 |
Quantum approximate optimization algorithm (QAOA) is one of the popular quantum algorithms that are used to solve combinatorial optimization problems via approximations. QAOA is able to be evaluated on both physical and virtual quantum computers simulated by classical computers, with virtual ones being favored for their noise-free feature and availability. Nevertheless, performing QAOA on virtual quantum computers suffers from a slow simulation speed for solving combinatorial optimization problems which require large-scale quantum circuit simulation (QCS). In this paper, we propose techniques to accelerate QCS for QAOA using mathematical optimizations to compress quantum operations, incorporating efficient bitwise operations to further lower the computational complexity, and leveraging different levels of parallelisms from modern multi-core processors, with a study case to show the effectiveness on solving max-cut problems.
|
cs
|
2,311.17792 |
At the Relativistic Heavy-Ion Collider (RHIC), the Polarized Atomic Hydrogen Gas Jet Target polarimeter (HJET) is employed for the precise measurement of the absolute transverse (vertical) polarization of proton beams, achieving low systematic uncertainties of approximately $\sigma^\text{syst}_P/P\leq0.5\%$. The acquired experimental data not only facilitated the determination of single $A_\text{N}(t)$ and double $A_\text{NN}(t)$ spin analyzing powers for 100 and 255 GeV proton beams but also revealed a non-zero Pomeron spin-flip contribution through a Regge fit. Preliminary results obtained for forward inelastic $p^{\uparrow}p$ and elastic $p^{\uparrow}A$ analyzing powers will be discussed. The success of HJET at RHIC suggests its potential application for proton beam polarimetry at the upcoming Electron-Ion Collider (EIC), aiming for an accuracy of 1\%. Moreover, the provided analysis indicates that the RHIC HJET target can serve as a tool for the precision calibration, with the required accuracy, of the $^3$He beam polarization at the EIC.
|
hep-ex
|
2,312.02748 |
Data-to-text generation involves transforming structured data, often represented as predicate-argument tuples, into coherent textual descriptions. Despite recent advances, systems still struggle when confronted with unseen combinations of predicates, producing unfaithful descriptions (e.g. hallucinations or omissions). We refer to this issue as compositional generalisation, and it encouraged us to create a benchmark for assessing the performance of different approaches on this specific problem. Furthermore, we propose a novel model that addresses compositional generalization by clustering predicates into groups. Our model generates text in a sentence-by-sentence manner, relying on one cluster of predicates at a time. This approach significantly outperforms T5~baselines across all evaluation metrics.Notably, it achieved a 31% improvement over T5 in terms of a metric focused on maintaining faithfulness to the input.
|
cs
|
2,312.04414 |
Integrated photonics has enabled countless technologies in telecommunications, spectroscopy, metrology, quantum optics, and quantum information processing. Using highly confined guided optical modes is the key that has made integrated circuits possible and has lead to scaling of complex designs, benefiting from their small footprint. At the same time, the field of quantum acoustics has recently gained significant attention due to its various potential advantages over its photonic counterparts, including smaller mode volume, lower energy, and orders of magnitude slower propagation speeds, as well as the potential for interconnecting distinct quantum systems. Developing analogous integrated phononic technology is critical for realizing the full potential of phonons and could lead to groundbreaking new applications, such as scalable quantum computing and hybrid quantum devices. In this work, we demonstrate for the first time a 4-port directional coupler for quantum mechanical excitations - a crucial component for integrated phononic circuits. Adjusting the length of the coupling region allows to realize phononic beam splitters with varying splitting ratios. By sending a single-phonon Fock state onto one of these phononic splitters, we demonstrate the capability of using the directional coupler directly in the quantum regime. Our work provides an essential step towards an integrated phononic platform for both classical and quantum technologies applications.
|
quant-ph
|
2,312.05698 |
In recent times, the field of unsupervised representation learning (URL) for time series data has garnered significant interest due to its remarkable adaptability across diverse downstream applications. Unsupervised learning goals differ from downstream tasks, making it tricky to ensure downstream task utility by focusing only on temporal feature characterization. Researchers have proposed multiple transformations to extract discriminative patterns implied in informative time series, trying to fill the gap. Despite the introduction of a variety of feature engineering techniques, e.g. spectral domain, wavelet transformed features, features in image form and symbolic features etc. the utilization of intricate feature fusion methods and dependence on heterogeneous features during inference hampers the scalability of the solutions. To address this, our study introduces an innovative approach that focuses on aligning and binding time series representations encoded from different modalities, inspired by spectral graph theory, thereby guiding the neural encoder to uncover latent pattern associations among these multi-modal features. In contrast to conventional methods that fuse features from multiple modalities, our proposed approach simplifies the neural architecture by retaining a single time series encoder, consequently leading to preserved scalability. We further demonstrate and prove mechanisms for the encoder to maintain better inductive bias. In our experimental evaluation, we validated the proposed method on a diverse set of time series datasets from various domains. Our approach outperforms existing state-of-the-art URL methods across diverse downstream tasks.
|
cs
|
2,312.06617 |
This paper is the first in a series of paper where we describe the differential operators on general nonlinear metric measure spaces, namely, the Finsler spaces. We try to propose a general method for gradient estimates of the positive solutions to nonlinear parabolic (or elliptic) equations on both compact and noncompact forward complete Finsler manifolds. The key of this approach is to find a valid comparison theorem, and particularly, to find related suitable curvature conditions. Actually, we define some non-Riemannian tensors and a generalization of the weighted Ricci curvature, called the mixed weighted Ricci curvature. With the assistance of these concepts, we prove a new kind of Laplacian comparison theorem. As an illustration of such method, we apply it to Finslerian Schrödinger equation. The Schrödinger equation is a typical class of parabolic equations, and the gradient estimation of their positive solution is representative. In this paper, based on the new Laplacian comparison theorem, we give global and local Li-Yau type gradient estimates for positive solutions to the Finslerian Schrödinger equation with respect to two different mixed weighted Ricci curvature conditions.
|
math
|
2,312.02336 |
We present six strongly gravitationally lensed Ly-$\alpha$ Emitters (LAEs) at $z\sim4-5$ with HST narrowband imaging isolating Ly-$\alpha$. Through complex radiative transfer Ly-$\alpha$ encodes information about the spatial distribution and kinematics of the neutral hydrogen upon which it scatters. We investigate the galaxy properties and Ly-$\alpha$ morphologies of our sample. Many previous studies of high-redshift LAEs have been limited in Ly-$\alpha$ spatial resolution. In this work we take advantage of high-resolution Ly-$\alpha$ imaging boosted by lensing magnification, allowing us to probe sub-galactic scales that are otherwise inaccessible at these redshifts. We use broadband imaging from HST (rest-frame UV) and Spitzer (rest-frame optical) in SED fitting; providing estimates on the stellar masses ($\sim 10^8 - 10^9 M_{\odot}$), stellar population ages ($t_{50} <40$ Myr), and amounts of dust ($A_V \sim 0.1 - 0.6$, statistically consistent with zero). We employ non-parametric star-formation histories to probe the young stellar-populations which create Ly-$\alpha$. We also examine the offsets between the Ly-$\alpha$ and stellar continuum, finding small upper limits of offsets ($< 0.1"$) consistent with studies of low-redshift LAEs; indicating our galaxies are not interacting or merging. Finally, we find a bimodality in our sample's Ly-$\alpha$ morphologies: clumpy and extended. We find a suggestive trend: our LAEs with clumpy Ly-$\alpha$ are generally younger than the LAEs with extended Ly-$\alpha$, suggesting a possible correlation with age.
|
astro-ph
|
2,312.04186 |
In the quest for fault-tolerant quantum computation using superconducting processors, accurate performance assessment and continuous design optimization stands at the forefront. To facilitate both meticulous simulation and streamlined design optimization, we introduce a multi-level simulation framework that spans both Hamiltonian and quantum error correction levels, and is equipped with the capability to compute gradients efficiently. This toolset aids in design optimization, tailored to specific objectives like quantum memory performance. Within our framework, we investigate the often-neglected spatially correlated unitary errors, highlighting their significant impact on logical error rates. We exemplify our approach through the multi-path coupling scheme of fluxonium qubits.
|
quant-ph
|
2,312.05568 |
The theory of Bayesian learning incorporates the use of Student-t Processes to model heavy-tailed distributions and datasets with outliers. However, despite Student-t Processes having a similar computational complexity as Gaussian Processes, there has been limited emphasis on the sparse representation of this model. This is mainly due to the increased difficulty in modeling and computation compared to previous sparse Gaussian Processes. Our motivation is to address the need for a sparse representation framework that reduces computational complexity, allowing Student-t Processes to be more flexible for real-world datasets. To achieve this, we leverage the conditional distribution of Student-t Processes to introduce sparse inducing points. Bayesian methods and variational inference are then utilized to derive a well-defined lower bound, facilitating more efficient optimization of our model through stochastic gradient descent. We propose two methods for computing the variational lower bound, one utilizing Monte Carlo sampling and the other employing Jensen's inequality to compute the KL regularization term in the loss function. We propose adopting these approaches as viable alternatives to Gaussian processes when the data might contain outliers or exhibit heavy-tailed behavior, and we provide specific recommendations for their applicability. We evaluate the two proposed approaches on various synthetic and real-world datasets from UCI and Kaggle, demonstrating their effectiveness compared to baseline methods in terms of computational complexity and accuracy, as well as their robustness to outliers.
|
cs
|
2,312.0256 |
In this note we give an upper bound on the Hausdorff dimension of removable setsfor elliptic and canceling homogeneous differential operators with constant coefficients in the class of bounded functions, using a simple extension of Frostman's lemma in Euclidean space with an additional power decay.
|
math
|
2,312.03226 |
Existing salient object detection methods are capable of predicting binary maps that highlight visually salient regions. However, these methods are limited in their ability to differentiate the relative importance of multiple objects and the relationships among them, which can lead to errors and reduced accuracy in downstream tasks that depend on the relative importance of multiple objects. To conquer, this paper proposes a new paradigm for saliency ranking, which aims to completely focus on ranking salient objects by their "importance order". While previous works have shown promising performance, they still face ill-posed problems. First, the saliency ranking ground truth (GT) orders generation methods are unreasonable since determining the correct ranking order is not well-defined, resulting in false alarms. Second, training a ranking model remains challenging because most saliency ranking methods follow the multi-task paradigm, leading to conflicts and trade-offs among different tasks. Third, existing regression-based saliency ranking methods are complex for saliency ranking models due to their reliance on instance mask-based saliency ranking orders. These methods require a significant amount of data to perform accurately and can be challenging to implement effectively. To solve these problems, this paper conducts an in-depth analysis of the causes and proposes a whole-flow processing paradigm of saliency ranking task from the perspective of "GT data generation", "network structure design" and "training protocol". The proposed approach outperforms existing state-of-the-art methods on the widely-used SALICON set, as demonstrated by extensive experiments with fair and reasonable comparisons. The saliency ranking task is still in its infancy, and our proposed unified framework can serve as a fundamental strategy to guide future work.
|
cs
|
2,312.04902 |
Deep neural networks (DNNs) are susceptible to backdoor attacks, where malicious functionality is embedded to allow attackers to trigger incorrect classifications. Old-school backdoor attacks use strong trigger features that can easily be learned by victim models. Despite robustness against input variation, the robustness however increases the likelihood of unintentional trigger activations. This leaves traces to existing defenses, which find approximate replacements for the original triggers that can activate the backdoor without being identical to the original trigger via, e.g., reverse engineering and sample overlay.
In this paper, we propose and investigate a new characteristic of backdoor attacks, namely, backdoor exclusivity, which measures the ability of backdoor triggers to remain effective in the presence of input variation. Building upon the concept of backdoor exclusivity, we propose Backdoor Exclusivity LifTing (BELT), a novel technique which suppresses the association between the backdoor and fuzzy triggers to enhance backdoor exclusivity for defense evasion. Extensive evaluation on three popular backdoor benchmarks validate, our approach substantially enhances the stealthiness of four old-school backdoor attacks, which, after backdoor exclusivity lifting, is able to evade six state-of-the-art backdoor countermeasures, at almost no cost of the attack success rate and normal utility. For example, one of the earliest backdoor attacks BadNet, enhanced by BELT, evades most of the state-of-the-art defenses including ABS and MOTH which would otherwise recognize the backdoored model.
|
cs
|
2,312.03213 |
Understanding model uncertainty is important for many applications. We propose Bootstrap Your Own Variance (BYOV), combining Bootstrap Your Own Latent (BYOL), a negative-free Self-Supervised Learning (SSL) algorithm, with Bayes by Backprop (BBB), a Bayesian method for estimating model posteriors. We find that the learned predictive std of BYOV vs. a supervised BBB model is well captured by a Gaussian distribution, providing preliminary evidence that the learned parameter posterior is useful for label free uncertainty estimation. BYOV improves upon the deterministic BYOL baseline (+2.83% test ECE, +1.03% test Brier) and presents better calibration and reliability when tested with various augmentations (eg: +2.4% test ECE, +1.2% test Brier for Salt & Pepper noise).
|
stat
|
2,312.03707 |
This study addresses the challenges of multi-label text classification. The difficulties arise from imbalanced data sets, varied text lengths, and numerous subjective feature labels. Existing solutions include traditional machine learning and deep neural networks for predictions. However, both approaches have their limitations. Traditional machine learning often overlooks the associations between words, while deep neural networks, despite their better classification performance, come with increased training complexity and time. This paper proposes a method utilizing the bag-of-words model approach based on the GloVe model and the CNN-BiLSTM network. The principle is to use the word vector matrix trained by the GloVe model as the input for the text embedding layer. Given that the GloVe model requires no further training, the neural network model can be trained more efficiently. The method achieves an accuracy rate of 87.26% on the test set and an F1 score of 0.8737, showcasing promising results.
|
cs
|
2,312.05386 |
Model extraction (ME) attacks represent one major threat to Machine-Learning-as-a-Service (MLaaS) platforms by ``stealing'' the functionality of confidential machine-learning models through querying black-box APIs. Over seven years have passed since ME attacks were first conceptualized in the seminal work. During this period, substantial advances have been made in both ME attacks and MLaaS platforms, raising the intriguing question: How has the vulnerability of MLaaS platforms to ME attacks been evolving? In this work, we conduct an in-depth study to answer this critical question. Specifically, we characterize the vulnerability of current, mainstream MLaaS platforms to ME attacks from multiple perspectives including attack strategies, learning techniques, surrogate-model design, and benchmark tasks. Many of our findings challenge previously reported results, suggesting emerging patterns of ME vulnerability. Further, by analyzing the vulnerability of the same MLaaS platforms using historical datasets from the past four years, we retrospectively characterize the evolution of ME vulnerability over time, leading to a set of interesting findings. Finally, we make suggestions about improving the current practice of MLaaS in terms of attack robustness. Our study sheds light on the current state of ME vulnerability in the wild and points to several promising directions for future research.
|
cs
|
2,312.0408 |
We present a theoretical study of resonance lifetimes in a two-component three-body system, specifically examining the decay of three-body resonances into a deep dimer and an unbound particle. Utilising the Gaussian expansion method together with the complex scaling method, we obtain the widths of these resonances from first principles. We focus on mass ratios in the typical range for mixtures of ultracold atoms and reveal a pronounced dependence of the resonance widths on the mass ratio: a distinct maximum near the equal-mass scenario and a rapid decrease away from it. Moreover, we show that this behaviour is not covered by the analytical formula of Pen'kov~[Phys. Rev. A 60, 3756 (1999)]. Notably, near the mass ratio for Caesium-Lithium mixtures, we obtain nearly vanishing widths of the resonances which validates to treat them in the bound state approximation. In addition, we perform our analysis on the resonance widths in both one and three dimensions and find that their qualitative dependence on the mass ratio agrees.
|
quant-ph
|
2,312.04616 |
Increasingly, there is well-grounded concern that through perpetual scaling-up of computation power and data, current deep learning techniques will create highly capable artificial intelligence that could pursue goals in a manner that is not aligned with human values. In turn, such AI could have the potential of leading to a scenario in which there is serious global-scale damage to human wellbeing. Against this backdrop, a number of researchers and public policy professionals have been developing ideas about how to govern AI in a manner that reduces the chances that it could lead to a global catastrophe. The jurisdictional focus of a vast majority of their assessments so far has been the United States, China, and Europe. That preference seems to reveal an assumption underlying most of the work in this field: That global south countries can only have a marginal role in attempts to govern AI development from a global catastrophic risk -focused perspective. Our paper sets out to undermine this assumption. We argue that global south countries like India and Singapore (and specific coalitions) could in fact be fairly consequential in the global catastrophic risk-focused governance of AI. We support our position using 4 key claims. 3 are constructed out of the current ways in which advanced foundational AI models are built and used while one is constructed on the strategic roles that global south countries and coalitions have historically played in the design and use of multilateral rules and institutions. As each claim is elaborated, we also suggest some ways through which global south countries can play a positive role in designing, strengthening and operationalizing global catastrophic risk-focused AI governance.
|
cs
|
2,312.02161 |
Due to 5G deployment, there is significant interest in LDPC decoding. While much research is devoted on efficient hardwiring of algorithms based on Belief Propagation (BP), it has been shown that LDPC decoding can be formulated as a combinatorial optimization problem, which could benefit from significant acceleration of physical computation mechanisms such as Ising machines. This approach has so far resulted in poor performance. This paper shows that the reason is not fundamental but suboptimal hardware and formulation. A co-designed Ising machine-based system can improve speed by 3 orders of magnitude. As a result, a physical computation approach can outperform hardwiring state-of-the-art algorithms. In this paper, we show such an augmented Ising machine that is 4.4$\times$ more energy efficient than the state of the art in the literature.
|
cs
|
2,312.06386 |
Monocular 3D human pose estimation (3D-HPE) is an inherently ambiguous task, as a 2D pose in an image might originate from different possible 3D poses. Yet, most 3D-HPE methods rely on regression models, which assume a one-to-one mapping between inputs and outputs. In this work, we provide theoretical and empirical evidence that, because of this ambiguity, common regression models are bound to predict topologically inconsistent poses, and that traditional evaluation metrics, such as the MPJPE, P-MPJPE and PCK, are insufficient to assess this aspect. As a solution, we propose ManiPose, a novel manifold-constrained multi-hypothesis model capable of proposing multiple candidate 3D poses for each 2D input, together with their corresponding plausibility. Unlike previous multi-hypothesis approaches, our solution is completely supervised and does not rely on complex generative models, thus greatly facilitating its training and usage. Furthermore, by constraining our model to lie within the human pose manifold, we can guarantee the consistency of all hypothetical poses predicted with our approach, which was not possible in previous works. We illustrate the usefulness of ManiPose in a synthetic 1D-to-2D lifting setting and demonstrate on real-world datasets that it outperforms state-of-the-art models in pose consistency by a large margin, while still reaching competitive MPJPE performance.
|
cs
|
2,312.03815 |
This paper envisions a revolutionary AIOS-Agent ecosystem, where Large Language Model (LLM) serves as the (Artificial) Intelligent Operating System (IOS, or AIOS)--an operating system "with soul". Upon this foundation, a diverse range of LLM-based AI Agent Applications (Agents, or AAPs) are developed, enriching the AIOS-Agent ecosystem and signaling a paradigm shift from the traditional OS-APP ecosystem. We envision that LLM's impact will not be limited to the AI application level, instead, it will in turn revolutionize the design and implementation of computer system, architecture, software, and programming language, featured by several main concepts: LLM as OS (system-level), Agents as Applications (application-level), Natural Language as Programming Interface (user-level), and Tools as Devices/Libraries (hardware/middleware-level). We begin by introducing the architecture of traditional OS. Then we formalize a conceptual framework for AIOS through "LLM as OS (LLMOS)", drawing analogies between AIOS and traditional OS: LLM is likened to OS kernel, context window to memory, external storage to file system, hardware tools to peripheral devices, software tools to programming libraries, and user prompts to user commands. Subsequently, we introduce the new AIOS-Agent Ecosystem, where users can easily program Agent Applications (AAPs) using natural language, democratizing the development of software, which is different from the traditional OS-APP ecosystem. Following this, we explore the diverse scope of Agent Applications. We delve into both single-agent and multi-agent systems, as well as human-agent interaction. Lastly, drawing on the insights from traditional OS-APP ecosystem, we propose a roadmap for the evolution of the AIOS-Agent ecosystem. This roadmap is designed to guide the future research and development, suggesting systematic progresses of AIOS and its Agent applications.
|
cs
|
2,312.04712 |
We present a method for identifying groups of test examples -- slices -- on which a model under-performs, a task now known as slice discovery. We formalize coherence -- a requirement that erroneous predictions, within a slice, should be wrong for the same reason -- as a key property that any slice discovery method should satisfy. We then use influence functions to derive a new slice discovery method, InfEmbed, which satisfies coherence by returning slices whose examples are influenced similarly by the training data. InfEmbed is simple, and consists of applying K-Means clustering to a novel representation we deem influence embeddings. We show InfEmbed outperforms current state-of-the-art methods on 2 benchmarks, and is effective for model debugging across several case studies.
|
cs
|
2,312.0335 |
Large and rich data is a prerequisite for effective training of deep neural networks. However, the irregularity of point cloud data makes manual annotation time-consuming and laborious. Self-supervised representation learning, which leverages the intrinsic structure of large-scale unlabelled data to learn meaningful feature representations, has attracted increasing attention in the field of point cloud research. However, self-supervised representation learning often suffers from model collapse, resulting in reduced information and diversity of the learned representation, and consequently degrading the performance of downstream tasks. To address this problem, we propose PointMoment, a novel framework for point cloud self-supervised representation learning that utilizes a high-order mixed moment loss function rather than the conventional contrastive loss function. Moreover, our framework does not require any special techniques such as asymmetric network architectures, gradient stopping, etc. Specifically, we calculate the high-order mixed moment of the feature variables and force them to decompose into products of their individual moment, thereby making multiple variables more independent and minimizing the feature redundancy. We also incorporate a contrastive learning approach to maximize the feature invariance under different data augmentations of the same point cloud. Experimental results show that our approach outperforms previous unsupervised learning methods on the downstream task of 3D point cloud classification and segmentation.
|
cs
|
2,312.05259 |
Due to the necessary security for the airport and flight, passengers are required to have strict security check before getting aboard. However, there are frequent complaints of wasting huge amount of time while waiting for the security check. This paper presents a potential solution aimed at optimizing gate setup procedures specifically tailored for Chicago OHare International Airport. By referring to queueing theory and performing Monte Carlo simulations, we propose an approach to significantly diminish the average waiting time to a more manageable level. Additionally, our study meticulously examines and identifies the influential factors contributing to this optimization, providing a comprehensive understanding of their impact.
|
cs
|
2,312.04581 |
In the present paper, we provide a detailed derivation of the stochastic Hamilton-Jacobi-Bellman equation
|
math
|
2,312.05045 |
Constraints on the quantum decoherence of entangled $\gamma$ quanta at the MeV scale, such as those produced following positron annihilation, have remained elusive for many decades. We present the first accurate, large acceptance experimental data for triple Compton scattering (TCS) of such entangled $\gamma$. An entanglement witness, the enhancement ($R$) of the azimuthal correlation between the final CS planes, is obtained where one of the $\gamma$ underwent intermediate Compton scattering (CS). $R$ is found to exceed the classical limit for scatter angles up to $\sim$40$^{\circ}$. The extracted $R$, over the range of the data ($0-70^{\circ}$), are better described by simple models in which entanglement is conserved or decohered proportional to the visibility of the scatter process. The work will help guide the development of a first theoretical description of decoherence at the MeV scale, a crucial next step for fundamental physics and entangled PET medical imaging.
|
quant-ph
|
2,312.04593 |
A network of chaotic systems can be designed in such a way that the cluster patterns formed by synchronous nodes can be controlled through the coupling parameters. We present a novel approach to exploit such a network for covert communication purposes, where controlled clusters encode the symbols spatio-temporally. The cluster synchronization network is divided into two subnetworks as transmitter and receiver. First, we specifically design the network whose controlled parameters reside in the transmitter. Second, we ensure that the nodes of the links connecting the transmitter and receiver are not in the same clusters for all the control parameters. The former condition ensures that the control parameters changed at the transmitter change the whole clustering scheme. The second condition enforces the transmitted signals are always continuous and chaotic. Hence, the transmitted signals are not modulated by the information directly, but distributed over the links connecting the subnetworks. The information cannot be deciphered by eavesdropping on the channel links without knowing the network topology. The performance has been assessed by extensive simulations of bit error rates under noisy channel conditions.
|
cs
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.