🌁#85: Curiosity, Open Source, and Timing: The Formula Behind DeepSeek’s Phenomenal Success
•
3
Ksenia Se
Kseniase
AI & ML interests
None yet
Recent Activity
upvoted
an
article
about 5 hours ago
🌁#85: Curiosity, Open Source, and Timing: The Formula Behind DeepSeek’s Phenomenal Success
upvoted
a
paper
about 7 hours ago
Evolution and The Knightian Blindspot of Machine Learning
reacted
to
their
post
with 👍
about 7 hours ago
7 Open-source Methods to Improve Video Generation and Understanding
AI community is making great strides toward achieving the full potential of multimodality in video generation and understanding. Last week studies showed that working with videos is now one of the main focuses for improving AI models. Another highlight of the week is that open source, once again, proves its value. For those who were impressed by DeepSeek-R1, we’re with you!
Today, we’re combining these two key focuses and bringing you a list of open-source methods for better video generation and understanding:
1. VideoLLaMA 3 model: Excels in various video and image tasks thanks to vision-centric training approach. https://huggingface.co/papers/2501.13106
2. FILMAGENT framework assigns roles to multiple AI agents, like a director, screenwriter, actor, and cinematographer, to automate the filmmaking process in 3D virtual environments. https://huggingface.co/papers/2501.12909
3. https://huggingface.co/papers/2501.13918 proposes a new VideoReward Model and approach that uses human feedback to refine video generation models.
4. DiffuEraser video inpainting model, based on stable diffusion, is designed to fill in missing areas with detailed, realistic content and to ensure consistent structures across frames. https://huggingface.co/papers/2501.10018
5. MAGI is a hybrid video gen model that combines masked and casual modeling. Its key innovation, Complete Teacher Forcing (CTF), conditions masked frames on fully visible frames. https://huggingface.co/papers/2501.12389
6. https://huggingface.co/papers/2501.08331 proposes motion control, allowing users to guide how objects or the camera move in generated videos. Its noise warping algorithm replaces random noise in videos with structured noise based on motion info.
7. Video Depth Anything model estimates depth consistently in super-long videos (several minutes or more) without sacrificing quality or speed. https://huggingface.co/papers/2501.12375
Articles
Organizations
Kseniase's activity
reacted to
their
post with 👍🚀
about 7 hours ago
Post
1888
7 Open-source Methods to Improve Video Generation and Understanding
AI community is making great strides toward achieving the full potential of multimodality in video generation and understanding. Last week studies showed that working with videos is now one of the main focuses for improving AI models. Another highlight of the week is that open source, once again, proves its value. For those who were impressed by DeepSeek-R1, we’re with you!
Today, we’re combining these two key focuses and bringing you a list of open-source methods for better video generation and understanding:
1. VideoLLaMA 3 model: Excels in various video and image tasks thanks to vision-centric training approach. VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understanding (2501.13106)
2. FILMAGENT framework assigns roles to multiple AI agents, like a director, screenwriter, actor, and cinematographer, to automate the filmmaking process in 3D virtual environments. FilmAgent: A Multi-Agent Framework for End-to-End Film Automation in Virtual 3D Spaces (2501.12909)
3. Improving Video Generation with Human Feedback (2501.13918) proposes a new VideoReward Model and approach that uses human feedback to refine video generation models.
4. DiffuEraser video inpainting model, based on stable diffusion, is designed to fill in missing areas with detailed, realistic content and to ensure consistent structures across frames. DiffuEraser: A Diffusion Model for Video Inpainting (2501.10018)
5. MAGI is a hybrid video gen model that combines masked and casual modeling. Its key innovation, Complete Teacher Forcing (CTF), conditions masked frames on fully visible frames. Taming Teacher Forcing for Masked Autoregressive Video Generation (2501.12389)
6. Go-with-the-Flow: Motion-Controllable Video Diffusion Models Using Real-Time Warped Noise (2501.08331) proposes motion control, allowing users to guide how objects or the camera move in generated videos. Its noise warping algorithm replaces random noise in videos with structured noise based on motion info.
7. Video Depth Anything model estimates depth consistently in super-long videos (several minutes or more) without sacrificing quality or speed. Video Depth Anything: Consistent Depth Estimation for Super-Long Videos (2501.12375)
AI community is making great strides toward achieving the full potential of multimodality in video generation and understanding. Last week studies showed that working with videos is now one of the main focuses for improving AI models. Another highlight of the week is that open source, once again, proves its value. For those who were impressed by DeepSeek-R1, we’re with you!
Today, we’re combining these two key focuses and bringing you a list of open-source methods for better video generation and understanding:
1. VideoLLaMA 3 model: Excels in various video and image tasks thanks to vision-centric training approach. VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understanding (2501.13106)
2. FILMAGENT framework assigns roles to multiple AI agents, like a director, screenwriter, actor, and cinematographer, to automate the filmmaking process in 3D virtual environments. FilmAgent: A Multi-Agent Framework for End-to-End Film Automation in Virtual 3D Spaces (2501.12909)
3. Improving Video Generation with Human Feedback (2501.13918) proposes a new VideoReward Model and approach that uses human feedback to refine video generation models.
4. DiffuEraser video inpainting model, based on stable diffusion, is designed to fill in missing areas with detailed, realistic content and to ensure consistent structures across frames. DiffuEraser: A Diffusion Model for Video Inpainting (2501.10018)
5. MAGI is a hybrid video gen model that combines masked and casual modeling. Its key innovation, Complete Teacher Forcing (CTF), conditions masked frames on fully visible frames. Taming Teacher Forcing for Masked Autoregressive Video Generation (2501.12389)
6. Go-with-the-Flow: Motion-Controllable Video Diffusion Models Using Real-Time Warped Noise (2501.08331) proposes motion control, allowing users to guide how objects or the camera move in generated videos. Its noise warping algorithm replaces random noise in videos with structured noise based on motion info.
7. Video Depth Anything model estimates depth consistently in super-long videos (several minutes or more) without sacrificing quality or speed. Video Depth Anything: Consistent Depth Estimation for Super-Long Videos (2501.12375)
reacted to
their
post with 🔥
1 day ago
Post
1888
7 Open-source Methods to Improve Video Generation and Understanding
AI community is making great strides toward achieving the full potential of multimodality in video generation and understanding. Last week studies showed that working with videos is now one of the main focuses for improving AI models. Another highlight of the week is that open source, once again, proves its value. For those who were impressed by DeepSeek-R1, we’re with you!
Today, we’re combining these two key focuses and bringing you a list of open-source methods for better video generation and understanding:
1. VideoLLaMA 3 model: Excels in various video and image tasks thanks to vision-centric training approach. VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understanding (2501.13106)
2. FILMAGENT framework assigns roles to multiple AI agents, like a director, screenwriter, actor, and cinematographer, to automate the filmmaking process in 3D virtual environments. FilmAgent: A Multi-Agent Framework for End-to-End Film Automation in Virtual 3D Spaces (2501.12909)
3. Improving Video Generation with Human Feedback (2501.13918) proposes a new VideoReward Model and approach that uses human feedback to refine video generation models.
4. DiffuEraser video inpainting model, based on stable diffusion, is designed to fill in missing areas with detailed, realistic content and to ensure consistent structures across frames. DiffuEraser: A Diffusion Model for Video Inpainting (2501.10018)
5. MAGI is a hybrid video gen model that combines masked and casual modeling. Its key innovation, Complete Teacher Forcing (CTF), conditions masked frames on fully visible frames. Taming Teacher Forcing for Masked Autoregressive Video Generation (2501.12389)
6. Go-with-the-Flow: Motion-Controllable Video Diffusion Models Using Real-Time Warped Noise (2501.08331) proposes motion control, allowing users to guide how objects or the camera move in generated videos. Its noise warping algorithm replaces random noise in videos with structured noise based on motion info.
7. Video Depth Anything model estimates depth consistently in super-long videos (several minutes or more) without sacrificing quality or speed. Video Depth Anything: Consistent Depth Estimation for Super-Long Videos (2501.12375)
AI community is making great strides toward achieving the full potential of multimodality in video generation and understanding. Last week studies showed that working with videos is now one of the main focuses for improving AI models. Another highlight of the week is that open source, once again, proves its value. For those who were impressed by DeepSeek-R1, we’re with you!
Today, we’re combining these two key focuses and bringing you a list of open-source methods for better video generation and understanding:
1. VideoLLaMA 3 model: Excels in various video and image tasks thanks to vision-centric training approach. VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understanding (2501.13106)
2. FILMAGENT framework assigns roles to multiple AI agents, like a director, screenwriter, actor, and cinematographer, to automate the filmmaking process in 3D virtual environments. FilmAgent: A Multi-Agent Framework for End-to-End Film Automation in Virtual 3D Spaces (2501.12909)
3. Improving Video Generation with Human Feedback (2501.13918) proposes a new VideoReward Model and approach that uses human feedback to refine video generation models.
4. DiffuEraser video inpainting model, based on stable diffusion, is designed to fill in missing areas with detailed, realistic content and to ensure consistent structures across frames. DiffuEraser: A Diffusion Model for Video Inpainting (2501.10018)
5. MAGI is a hybrid video gen model that combines masked and casual modeling. Its key innovation, Complete Teacher Forcing (CTF), conditions masked frames on fully visible frames. Taming Teacher Forcing for Masked Autoregressive Video Generation (2501.12389)
6. Go-with-the-Flow: Motion-Controllable Video Diffusion Models Using Real-Time Warped Noise (2501.08331) proposes motion control, allowing users to guide how objects or the camera move in generated videos. Its noise warping algorithm replaces random noise in videos with structured noise based on motion info.
7. Video Depth Anything model estimates depth consistently in super-long videos (several minutes or more) without sacrificing quality or speed. Video Depth Anything: Consistent Depth Estimation for Super-Long Videos (2501.12375)
posted
an
update
1 day ago
Post
1888
7 Open-source Methods to Improve Video Generation and Understanding
AI community is making great strides toward achieving the full potential of multimodality in video generation and understanding. Last week studies showed that working with videos is now one of the main focuses for improving AI models. Another highlight of the week is that open source, once again, proves its value. For those who were impressed by DeepSeek-R1, we’re with you!
Today, we’re combining these two key focuses and bringing you a list of open-source methods for better video generation and understanding:
1. VideoLLaMA 3 model: Excels in various video and image tasks thanks to vision-centric training approach. VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understanding (2501.13106)
2. FILMAGENT framework assigns roles to multiple AI agents, like a director, screenwriter, actor, and cinematographer, to automate the filmmaking process in 3D virtual environments. FilmAgent: A Multi-Agent Framework for End-to-End Film Automation in Virtual 3D Spaces (2501.12909)
3. Improving Video Generation with Human Feedback (2501.13918) proposes a new VideoReward Model and approach that uses human feedback to refine video generation models.
4. DiffuEraser video inpainting model, based on stable diffusion, is designed to fill in missing areas with detailed, realistic content and to ensure consistent structures across frames. DiffuEraser: A Diffusion Model for Video Inpainting (2501.10018)
5. MAGI is a hybrid video gen model that combines masked and casual modeling. Its key innovation, Complete Teacher Forcing (CTF), conditions masked frames on fully visible frames. Taming Teacher Forcing for Masked Autoregressive Video Generation (2501.12389)
6. Go-with-the-Flow: Motion-Controllable Video Diffusion Models Using Real-Time Warped Noise (2501.08331) proposes motion control, allowing users to guide how objects or the camera move in generated videos. Its noise warping algorithm replaces random noise in videos with structured noise based on motion info.
7. Video Depth Anything model estimates depth consistently in super-long videos (several minutes or more) without sacrificing quality or speed. Video Depth Anything: Consistent Depth Estimation for Super-Long Videos (2501.12375)
AI community is making great strides toward achieving the full potential of multimodality in video generation and understanding. Last week studies showed that working with videos is now one of the main focuses for improving AI models. Another highlight of the week is that open source, once again, proves its value. For those who were impressed by DeepSeek-R1, we’re with you!
Today, we’re combining these two key focuses and bringing you a list of open-source methods for better video generation and understanding:
1. VideoLLaMA 3 model: Excels in various video and image tasks thanks to vision-centric training approach. VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understanding (2501.13106)
2. FILMAGENT framework assigns roles to multiple AI agents, like a director, screenwriter, actor, and cinematographer, to automate the filmmaking process in 3D virtual environments. FilmAgent: A Multi-Agent Framework for End-to-End Film Automation in Virtual 3D Spaces (2501.12909)
3. Improving Video Generation with Human Feedback (2501.13918) proposes a new VideoReward Model and approach that uses human feedback to refine video generation models.
4. DiffuEraser video inpainting model, based on stable diffusion, is designed to fill in missing areas with detailed, realistic content and to ensure consistent structures across frames. DiffuEraser: A Diffusion Model for Video Inpainting (2501.10018)
5. MAGI is a hybrid video gen model that combines masked and casual modeling. Its key innovation, Complete Teacher Forcing (CTF), conditions masked frames on fully visible frames. Taming Teacher Forcing for Masked Autoregressive Video Generation (2501.12389)
6. Go-with-the-Flow: Motion-Controllable Video Diffusion Models Using Real-Time Warped Noise (2501.08331) proposes motion control, allowing users to guide how objects or the camera move in generated videos. Its noise warping algorithm replaces random noise in videos with structured noise based on motion info.
7. Video Depth Anything model estimates depth consistently in super-long videos (several minutes or more) without sacrificing quality or speed. Video Depth Anything: Consistent Depth Estimation for Super-Long Videos (2501.12375)
reacted to
their
post with 👀
7 days ago
Post
1859
10 Recent Advancements in Math Reasoning
Over the last few weeks, we have witnessed a surge in AI models' math reasoning capabilities. Top companies like Microsoft, NVIDIA, and Alibaba Qwen have already joined this race to make models "smarter" in mathematics. But why is this shift happening now?
Complex math calculations require advanced multi-step reasoning, making mathematics an ideal domain for demonstrating a model's strong "thinking" capabilities. Additionally, as AI continues to evolve and is applied in math-intensive fields such as machine learning and quantum computing (which is predicted to see significant growth in 2025), it must meet the demands of complex reasoning.
Moreover, AI models can be integrated with external tools like symbolic solvers or computational engines to tackle large-scale math problems, which also needs high-quality math reasoning.
So here’s a list of 10 recent advancements in math reasoning of AI models:
1. NVIDIA: AceMath: Advancing Frontier Math Reasoning with Post-Training and Reward Modeling (2412.15084)
2. Qwen, Alibaba: Qwen2.5-Math-PRM The Lessons of Developing Process Reward Models in Mathematical Reasoning (2501.07301) and PROCESSBENCH evaluation ProcessBench: Identifying Process Errors in Mathematical Reasoning (2412.06559)
3. Microsoft Research: rStar-Math: Small LLMs Can Master Math Reasoning with Self-Evolved Deep Thinking (2501.04519)
4. BoostStep: Boosting mathematical capability of Large Language Models via improved single-step reasoning (2501.03226)
5. URSA: Understanding and Verifying Chain-of-thought Reasoning in Multimodal Mathematics (2501.04686)
6. U-MATH: A University-Level Benchmark for Evaluating Mathematical Skills in LLMs (2412.03205)
7. Open Eyes, Then Reason: Fine-grained Visual Mathematical Understanding in MLLMs (2501.06430)
8. End-to-End Bangla AI for Solving Math Olympiad Problem Benchmark: Leveraging Large Language Model Using Integrated Approach (2501.04425)
9. Quantization Meets Reasoning: Exploring LLM Low-Bit Quantization Degradation for Mathematical Reasoning (2501.03035)
10. System-2 Mathematical Reasoning via Enriched Instruction Tuning (2412.16964)
Over the last few weeks, we have witnessed a surge in AI models' math reasoning capabilities. Top companies like Microsoft, NVIDIA, and Alibaba Qwen have already joined this race to make models "smarter" in mathematics. But why is this shift happening now?
Complex math calculations require advanced multi-step reasoning, making mathematics an ideal domain for demonstrating a model's strong "thinking" capabilities. Additionally, as AI continues to evolve and is applied in math-intensive fields such as machine learning and quantum computing (which is predicted to see significant growth in 2025), it must meet the demands of complex reasoning.
Moreover, AI models can be integrated with external tools like symbolic solvers or computational engines to tackle large-scale math problems, which also needs high-quality math reasoning.
So here’s a list of 10 recent advancements in math reasoning of AI models:
1. NVIDIA: AceMath: Advancing Frontier Math Reasoning with Post-Training and Reward Modeling (2412.15084)
2. Qwen, Alibaba: Qwen2.5-Math-PRM The Lessons of Developing Process Reward Models in Mathematical Reasoning (2501.07301) and PROCESSBENCH evaluation ProcessBench: Identifying Process Errors in Mathematical Reasoning (2412.06559)
3. Microsoft Research: rStar-Math: Small LLMs Can Master Math Reasoning with Self-Evolved Deep Thinking (2501.04519)
4. BoostStep: Boosting mathematical capability of Large Language Models via improved single-step reasoning (2501.03226)
5. URSA: Understanding and Verifying Chain-of-thought Reasoning in Multimodal Mathematics (2501.04686)
6. U-MATH: A University-Level Benchmark for Evaluating Mathematical Skills in LLMs (2412.03205)
7. Open Eyes, Then Reason: Fine-grained Visual Mathematical Understanding in MLLMs (2501.06430)
8. End-to-End Bangla AI for Solving Math Olympiad Problem Benchmark: Leveraging Large Language Model Using Integrated Approach (2501.04425)
9. Quantization Meets Reasoning: Exploring LLM Low-Bit Quantization Degradation for Mathematical Reasoning (2501.03035)
10. System-2 Mathematical Reasoning via Enriched Instruction Tuning (2412.16964)
posted
an
update
9 days ago
Post
1859
10 Recent Advancements in Math Reasoning
Over the last few weeks, we have witnessed a surge in AI models' math reasoning capabilities. Top companies like Microsoft, NVIDIA, and Alibaba Qwen have already joined this race to make models "smarter" in mathematics. But why is this shift happening now?
Complex math calculations require advanced multi-step reasoning, making mathematics an ideal domain for demonstrating a model's strong "thinking" capabilities. Additionally, as AI continues to evolve and is applied in math-intensive fields such as machine learning and quantum computing (which is predicted to see significant growth in 2025), it must meet the demands of complex reasoning.
Moreover, AI models can be integrated with external tools like symbolic solvers or computational engines to tackle large-scale math problems, which also needs high-quality math reasoning.
So here’s a list of 10 recent advancements in math reasoning of AI models:
1. NVIDIA: AceMath: Advancing Frontier Math Reasoning with Post-Training and Reward Modeling (2412.15084)
2. Qwen, Alibaba: Qwen2.5-Math-PRM The Lessons of Developing Process Reward Models in Mathematical Reasoning (2501.07301) and PROCESSBENCH evaluation ProcessBench: Identifying Process Errors in Mathematical Reasoning (2412.06559)
3. Microsoft Research: rStar-Math: Small LLMs Can Master Math Reasoning with Self-Evolved Deep Thinking (2501.04519)
4. BoostStep: Boosting mathematical capability of Large Language Models via improved single-step reasoning (2501.03226)
5. URSA: Understanding and Verifying Chain-of-thought Reasoning in Multimodal Mathematics (2501.04686)
6. U-MATH: A University-Level Benchmark for Evaluating Mathematical Skills in LLMs (2412.03205)
7. Open Eyes, Then Reason: Fine-grained Visual Mathematical Understanding in MLLMs (2501.06430)
8. End-to-End Bangla AI for Solving Math Olympiad Problem Benchmark: Leveraging Large Language Model Using Integrated Approach (2501.04425)
9. Quantization Meets Reasoning: Exploring LLM Low-Bit Quantization Degradation for Mathematical Reasoning (2501.03035)
10. System-2 Mathematical Reasoning via Enriched Instruction Tuning (2412.16964)
Over the last few weeks, we have witnessed a surge in AI models' math reasoning capabilities. Top companies like Microsoft, NVIDIA, and Alibaba Qwen have already joined this race to make models "smarter" in mathematics. But why is this shift happening now?
Complex math calculations require advanced multi-step reasoning, making mathematics an ideal domain for demonstrating a model's strong "thinking" capabilities. Additionally, as AI continues to evolve and is applied in math-intensive fields such as machine learning and quantum computing (which is predicted to see significant growth in 2025), it must meet the demands of complex reasoning.
Moreover, AI models can be integrated with external tools like symbolic solvers or computational engines to tackle large-scale math problems, which also needs high-quality math reasoning.
So here’s a list of 10 recent advancements in math reasoning of AI models:
1. NVIDIA: AceMath: Advancing Frontier Math Reasoning with Post-Training and Reward Modeling (2412.15084)
2. Qwen, Alibaba: Qwen2.5-Math-PRM The Lessons of Developing Process Reward Models in Mathematical Reasoning (2501.07301) and PROCESSBENCH evaluation ProcessBench: Identifying Process Errors in Mathematical Reasoning (2412.06559)
3. Microsoft Research: rStar-Math: Small LLMs Can Master Math Reasoning with Self-Evolved Deep Thinking (2501.04519)
4. BoostStep: Boosting mathematical capability of Large Language Models via improved single-step reasoning (2501.03226)
5. URSA: Understanding and Verifying Chain-of-thought Reasoning in Multimodal Mathematics (2501.04686)
6. U-MATH: A University-Level Benchmark for Evaluating Mathematical Skills in LLMs (2412.03205)
7. Open Eyes, Then Reason: Fine-grained Visual Mathematical Understanding in MLLMs (2501.06430)
8. End-to-End Bangla AI for Solving Math Olympiad Problem Benchmark: Leveraging Large Language Model Using Integrated Approach (2501.04425)
9. Quantization Meets Reasoning: Exploring LLM Low-Bit Quantization Degradation for Mathematical Reasoning (2501.03035)
10. System-2 Mathematical Reasoning via Enriched Instruction Tuning (2412.16964)
posted
an
update
10 days ago
Post
694
Today, we spoke with Snowflake’s AI Research Team Leads, Yuxiong He and Samyam Rajbhandari (
@samyam
) (he is also one the researchers behind
DeepSpeed-FastGen: High-throughput Text Generation for LLMs via MII and
DeepSpeed-Inference (2401.08671) and other DeepSpeed papers)
Collaborating with their co-authors to reduce inference costs for enterprise-specific tasks, they observed that inputs are often significantly larger than outputs. This is because it’s in the nature of enterprises to analyze enormous amounts of information trying to extract valuable insights, which are much shorter. To address this, they developed SwiftKV SwiftKV: Fast Prefill-Optimized Inference with Knowledge-Preserving Model Transformation (2410.03960), an optimization that reduces LLM inference costs by up to 75% for Meta Llama LLMs, enhancing efficiency and performance in enterprise AI tasks.
Today they are open-sourcing SwiftKV ( Snowflake/Llama-3.1-SwiftKV-8B-Instruct) and ArcticTrainging Platform.
In our new episode "15 minutes with a Researcher" they explain how SwiftKV works, its applicability to other architectures, its limitations, and additional methods to further reduce computation costs in inference.
Watch the full 15 min interview here (https://youtu.be/9x1k7eXe-6Q?si=4_HQOyi1CPHgvlrx)
Collaborating with their co-authors to reduce inference costs for enterprise-specific tasks, they observed that inputs are often significantly larger than outputs. This is because it’s in the nature of enterprises to analyze enormous amounts of information trying to extract valuable insights, which are much shorter. To address this, they developed SwiftKV SwiftKV: Fast Prefill-Optimized Inference with Knowledge-Preserving Model Transformation (2410.03960), an optimization that reduces LLM inference costs by up to 75% for Meta Llama LLMs, enhancing efficiency and performance in enterprise AI tasks.
Today they are open-sourcing SwiftKV ( Snowflake/Llama-3.1-SwiftKV-8B-Instruct) and ArcticTrainging Platform.
In our new episode "15 minutes with a Researcher" they explain how SwiftKV works, its applicability to other architectures, its limitations, and additional methods to further reduce computation costs in inference.
Watch the full 15 min interview here (https://youtu.be/9x1k7eXe-6Q?si=4_HQOyi1CPHgvlrx)
reacted to
their
post with 🔥
14 days ago
Post
1860
10 AI Systems for Scientific Research
Almost every AI researcher has studied or conducted a large number of AI research papers. So, it's quite logical that researchers are trying to create AI systems to help conduct research. Creating scientific research could be much easier and more varied if we use LLMs and AI assistants tailored for this purpose. Just imagine how interesting it would be to read high-quality research about AI made by an AI agent.
Today, we offer you to explore these 10 AI systems for scientific research:
1. Agent Laboratory framework helps researchers input their ideas by generating a research report and code repository: Agent Laboratory: Using LLM Agents as Research Assistants (2501.04227)
2. AI Scientist performs fully automated scientific discovery including creating ideas: The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery (2408.06292)
3. SciMON generates new ideas derived from the scientific literature: Learning to Generate Novel Scientific Directions with Contextualized Literature-based Discovery (2305.14259)
4. ResearchAgent implements LLMs to automate idea generation, methods, and experiment design, and ReviewingAgents' feedback to refine ideas: ResearchAgent: Iterative Research Idea Generation over Scientific Literature with Large Language Models (2404.07738)
5. Scientific Generative Agent (SGA) discovers novel, coherent solutions in physics and molecular design: LLM and Simulation as Bilevel Optimizers: A New Paradigm to Advance Physical Scientific Discovery (2405.09783)
6. MLRCopilot boosts machine learning research: MLR-Copilot: Autonomous Machine Learning Research based on Large Language Models Agents (2408.14033)
7. SciAgents accelerates material science discovery through combining knowledge graphs, LLMs, and multi-agent systems. SciAgents: Automating scientific discovery through multi-agent intelligent graph reasoning (2409.05556)
8. VirSci multi-agent system mimics teamwork among scientists. Two Heads Are Better Than One: A Multi-Agent System Has the Potential to Improve Scientific Idea Generation (2410.09403)
9. Chain-of-Ideas (CoI) agent organizes research into a chain structure. Chain of Ideas: Revolutionizing Research in Novel Idea Development with LLM Agents (2410.13185)
10. A system with CycleResearcher and CycleReviewer generates research papers and peer reviews: CycleResearcher: Improving Automated Research via Automated Review (2411.00816)
LLM4SR: A Survey on Large Language Models for Scientific Research (2501.04306) is worth exploring to study and analyze more systems for scientific research
Almost every AI researcher has studied or conducted a large number of AI research papers. So, it's quite logical that researchers are trying to create AI systems to help conduct research. Creating scientific research could be much easier and more varied if we use LLMs and AI assistants tailored for this purpose. Just imagine how interesting it would be to read high-quality research about AI made by an AI agent.
Today, we offer you to explore these 10 AI systems for scientific research:
1. Agent Laboratory framework helps researchers input their ideas by generating a research report and code repository: Agent Laboratory: Using LLM Agents as Research Assistants (2501.04227)
2. AI Scientist performs fully automated scientific discovery including creating ideas: The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery (2408.06292)
3. SciMON generates new ideas derived from the scientific literature: Learning to Generate Novel Scientific Directions with Contextualized Literature-based Discovery (2305.14259)
4. ResearchAgent implements LLMs to automate idea generation, methods, and experiment design, and ReviewingAgents' feedback to refine ideas: ResearchAgent: Iterative Research Idea Generation over Scientific Literature with Large Language Models (2404.07738)
5. Scientific Generative Agent (SGA) discovers novel, coherent solutions in physics and molecular design: LLM and Simulation as Bilevel Optimizers: A New Paradigm to Advance Physical Scientific Discovery (2405.09783)
6. MLRCopilot boosts machine learning research: MLR-Copilot: Autonomous Machine Learning Research based on Large Language Models Agents (2408.14033)
7. SciAgents accelerates material science discovery through combining knowledge graphs, LLMs, and multi-agent systems. SciAgents: Automating scientific discovery through multi-agent intelligent graph reasoning (2409.05556)
8. VirSci multi-agent system mimics teamwork among scientists. Two Heads Are Better Than One: A Multi-Agent System Has the Potential to Improve Scientific Idea Generation (2410.09403)
9. Chain-of-Ideas (CoI) agent organizes research into a chain structure. Chain of Ideas: Revolutionizing Research in Novel Idea Development with LLM Agents (2410.13185)
10. A system with CycleResearcher and CycleReviewer generates research papers and peer reviews: CycleResearcher: Improving Automated Research via Automated Review (2411.00816)
LLM4SR: A Survey on Large Language Models for Scientific Research (2501.04306) is worth exploring to study and analyze more systems for scientific research
posted
an
update
16 days ago
Post
1860
10 AI Systems for Scientific Research
Almost every AI researcher has studied or conducted a large number of AI research papers. So, it's quite logical that researchers are trying to create AI systems to help conduct research. Creating scientific research could be much easier and more varied if we use LLMs and AI assistants tailored for this purpose. Just imagine how interesting it would be to read high-quality research about AI made by an AI agent.
Today, we offer you to explore these 10 AI systems for scientific research:
1. Agent Laboratory framework helps researchers input their ideas by generating a research report and code repository: Agent Laboratory: Using LLM Agents as Research Assistants (2501.04227)
2. AI Scientist performs fully automated scientific discovery including creating ideas: The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery (2408.06292)
3. SciMON generates new ideas derived from the scientific literature: Learning to Generate Novel Scientific Directions with Contextualized Literature-based Discovery (2305.14259)
4. ResearchAgent implements LLMs to automate idea generation, methods, and experiment design, and ReviewingAgents' feedback to refine ideas: ResearchAgent: Iterative Research Idea Generation over Scientific Literature with Large Language Models (2404.07738)
5. Scientific Generative Agent (SGA) discovers novel, coherent solutions in physics and molecular design: LLM and Simulation as Bilevel Optimizers: A New Paradigm to Advance Physical Scientific Discovery (2405.09783)
6. MLRCopilot boosts machine learning research: MLR-Copilot: Autonomous Machine Learning Research based on Large Language Models Agents (2408.14033)
7. SciAgents accelerates material science discovery through combining knowledge graphs, LLMs, and multi-agent systems. SciAgents: Automating scientific discovery through multi-agent intelligent graph reasoning (2409.05556)
8. VirSci multi-agent system mimics teamwork among scientists. Two Heads Are Better Than One: A Multi-Agent System Has the Potential to Improve Scientific Idea Generation (2410.09403)
9. Chain-of-Ideas (CoI) agent organizes research into a chain structure. Chain of Ideas: Revolutionizing Research in Novel Idea Development with LLM Agents (2410.13185)
10. A system with CycleResearcher and CycleReviewer generates research papers and peer reviews: CycleResearcher: Improving Automated Research via Automated Review (2411.00816)
LLM4SR: A Survey on Large Language Models for Scientific Research (2501.04306) is worth exploring to study and analyze more systems for scientific research
Almost every AI researcher has studied or conducted a large number of AI research papers. So, it's quite logical that researchers are trying to create AI systems to help conduct research. Creating scientific research could be much easier and more varied if we use LLMs and AI assistants tailored for this purpose. Just imagine how interesting it would be to read high-quality research about AI made by an AI agent.
Today, we offer you to explore these 10 AI systems for scientific research:
1. Agent Laboratory framework helps researchers input their ideas by generating a research report and code repository: Agent Laboratory: Using LLM Agents as Research Assistants (2501.04227)
2. AI Scientist performs fully automated scientific discovery including creating ideas: The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery (2408.06292)
3. SciMON generates new ideas derived from the scientific literature: Learning to Generate Novel Scientific Directions with Contextualized Literature-based Discovery (2305.14259)
4. ResearchAgent implements LLMs to automate idea generation, methods, and experiment design, and ReviewingAgents' feedback to refine ideas: ResearchAgent: Iterative Research Idea Generation over Scientific Literature with Large Language Models (2404.07738)
5. Scientific Generative Agent (SGA) discovers novel, coherent solutions in physics and molecular design: LLM and Simulation as Bilevel Optimizers: A New Paradigm to Advance Physical Scientific Discovery (2405.09783)
6. MLRCopilot boosts machine learning research: MLR-Copilot: Autonomous Machine Learning Research based on Large Language Models Agents (2408.14033)
7. SciAgents accelerates material science discovery through combining knowledge graphs, LLMs, and multi-agent systems. SciAgents: Automating scientific discovery through multi-agent intelligent graph reasoning (2409.05556)
8. VirSci multi-agent system mimics teamwork among scientists. Two Heads Are Better Than One: A Multi-Agent System Has the Potential to Improve Scientific Idea Generation (2410.09403)
9. Chain-of-Ideas (CoI) agent organizes research into a chain structure. Chain of Ideas: Revolutionizing Research in Novel Idea Development with LLM Agents (2410.13185)
10. A system with CycleResearcher and CycleReviewer generates research papers and peer reviews: CycleResearcher: Improving Automated Research via Automated Review (2411.00816)
LLM4SR: A Survey on Large Language Models for Scientific Research (2501.04306) is worth exploring to study and analyze more systems for scientific research
reacted to
Severian's
post with 👍
19 days ago
Post
3852
Interesting Solution to the Problem of Misguided Attention
So I've been fascinated by the problem of Misguided Attention for a few weeks. I am trying to build an inference algorithm to help LLMs address that issue; but in the process, I found a cool short-term fix I call "Mindful Attention" using just prompt-engineering.
Have you ever thought about how our brains filter reality through layers of past experiences, concepts, and mental images? For example, when you look at an oak tree, are you truly seeing that oak tree in all its unique details, or are you overlaying it with a generalized idea of "oak tree"? This phenomenon inspired the new approach.
LLMs often fall into a similar trap, hence the Misguided Attention problem. They process input not as it’s uniquely presented but through patterns and templates they’ve seen before. This leads to responses that can feel "off," like missing the point of a carefully crafted prompt or defaulting to familiar but irrelevant solutions.
I wanted to address this head-on by encouraging LLMs to slow down, focus, and engage directly with the input—free of assumptions. This is the core of the Mindful Attention Directive, a prompt designed to steer models away from over-generalization and back into the moment.
You can read more about the broader issue here: https://github.com/cpldcpu/MisguidedAttention
And if you want to try this mindful approach in action, check out the LLM I’ve set up for testing: https://hf.co/chat/assistant/677e7ebcb0f26b87340f032e. It works about 80% of the time to counteract these issues, and the results are pretty cool.
I'll add the Gist with the full prompt. I admit, it is quite verbose but it's the most effective one I have landed on yet. I am working on a smaller version that can be appended to any System Prompt to harness the Mindful Attention. Feel free to experiment to find a better version for the community!
Here is the Gist: https://gist.github.com/severian42/6dd96a94e546a38642278aeb4537cfb3
So I've been fascinated by the problem of Misguided Attention for a few weeks. I am trying to build an inference algorithm to help LLMs address that issue; but in the process, I found a cool short-term fix I call "Mindful Attention" using just prompt-engineering.
Have you ever thought about how our brains filter reality through layers of past experiences, concepts, and mental images? For example, when you look at an oak tree, are you truly seeing that oak tree in all its unique details, or are you overlaying it with a generalized idea of "oak tree"? This phenomenon inspired the new approach.
LLMs often fall into a similar trap, hence the Misguided Attention problem. They process input not as it’s uniquely presented but through patterns and templates they’ve seen before. This leads to responses that can feel "off," like missing the point of a carefully crafted prompt or defaulting to familiar but irrelevant solutions.
I wanted to address this head-on by encouraging LLMs to slow down, focus, and engage directly with the input—free of assumptions. This is the core of the Mindful Attention Directive, a prompt designed to steer models away from over-generalization and back into the moment.
You can read more about the broader issue here: https://github.com/cpldcpu/MisguidedAttention
And if you want to try this mindful approach in action, check out the LLM I’ve set up for testing: https://hf.co/chat/assistant/677e7ebcb0f26b87340f032e. It works about 80% of the time to counteract these issues, and the results are pretty cool.
I'll add the Gist with the full prompt. I admit, it is quite verbose but it's the most effective one I have landed on yet. I am working on a smaller version that can be appended to any System Prompt to harness the Mindful Attention. Feel free to experiment to find a better version for the community!
Here is the Gist: https://gist.github.com/severian42/6dd96a94e546a38642278aeb4537cfb3
reacted to
merve's
post with ❤️
19 days ago
Post
4830
supercharge your LLM apps with smolagents 🔥
however cool your LLM is, without being agentic it can only go so far
enter smolagents: a new agent library by Hugging Face to make the LLM write code, do analysis and automate boring stuff!
Here's our blog for you to get started https://huggingface.co/blog/smolagents
however cool your LLM is, without being agentic it can only go so far
enter smolagents: a new agent library by Hugging Face to make the LLM write code, do analysis and automate boring stuff!
Here's our blog for you to get started https://huggingface.co/blog/smolagents
reacted to
m-ric's
post with ❤️
19 days ago
Post
5055
Since I published it on GitHub a few days ago,
Hugging Face's new agentic library 𝘀𝗺𝗼𝗹𝗮𝗴𝗲𝗻𝘁𝘀 has gathered nearly 4k stars 🤯
➡️ But we are just getting started on agents: so we are hiring an ML Engineer to join me and double down on this effort!
The plan is to build GUI agents: agents that can act on your computer with mouse & keyboard, like Claude Computer Use.
We will make it work better, and fully open. ✨
Sounds like something you'd like to do? Apply here 👉 https://apply.workable.com/huggingface/j/AF1D4E3FEB/
Hugging Face's new agentic library 𝘀𝗺𝗼𝗹𝗮𝗴𝗲𝗻𝘁𝘀 has gathered nearly 4k stars 🤯
➡️ But we are just getting started on agents: so we are hiring an ML Engineer to join me and double down on this effort!
The plan is to build GUI agents: agents that can act on your computer with mouse & keyboard, like Claude Computer Use.
We will make it work better, and fully open. ✨
Sounds like something you'd like to do? Apply here 👉 https://apply.workable.com/huggingface/j/AF1D4E3FEB/
- 3 replies
posted
an
update
30 days ago
Post
2588
10 Free Comprehensive Datasets for Supervised Fine-Tuning
High-quality datasets, their size and relevance directly impact the effectiveness of fine-tuning and the models' real-world applications. Among the numerous datasets for different tasks, it can be challenging to choose the most comprehensive dataset that best suits your purposes.
So today, we invite you to explore top 10 free datasets on natural language processing and maths:
1. fka/awesome-chatgpt-prompts proposes a huge variety of prompts that can be used with ChatGPT. Over 700 models were trained on this dataset.
2. HuggingFaceFW/fineweb from Hugging Face includes 15T tokens of cleaned and deduplicated English web data. It’s suitable for LLM training, benchmarking, model validation.
3. HuggingFaceFW/fineweb-2 is an another version of FineWeb with high-quality pretraining data to over 1000 languages.
4. O1-OPEN/OpenO1-SFT with Chinese and English data can be used for Chain-of-Thought activation.
5. yahma/alpaca-cleaned is a curated version of the original Alpaca Dataset released by Stanford.
6. lmsys/lmsys-chat-1m with 1 million real-world conversations with 25 state-of-the-art LLMs offers diverse use cases, like content moderation, safety benchmarks, and training instruction-following models.
7. allenai/dolma from Allen AI includes 3T tokens from a diverse mix of web content, academic publications, code, books, and encyclopedic materials.
Math datasets:
1. HuggingFaceTB/finemath consists of educational math content and has two versions: 34B tokens and 54B tokens.
2. amphora/QwQ-LongCoT-130K for training O1-like LLMs.
3. openai/gsm8k for training multi-step reasoning.
High-quality datasets, their size and relevance directly impact the effectiveness of fine-tuning and the models' real-world applications. Among the numerous datasets for different tasks, it can be challenging to choose the most comprehensive dataset that best suits your purposes.
So today, we invite you to explore top 10 free datasets on natural language processing and maths:
1. fka/awesome-chatgpt-prompts proposes a huge variety of prompts that can be used with ChatGPT. Over 700 models were trained on this dataset.
2. HuggingFaceFW/fineweb from Hugging Face includes 15T tokens of cleaned and deduplicated English web data. It’s suitable for LLM training, benchmarking, model validation.
3. HuggingFaceFW/fineweb-2 is an another version of FineWeb with high-quality pretraining data to over 1000 languages.
4. O1-OPEN/OpenO1-SFT with Chinese and English data can be used for Chain-of-Thought activation.
5. yahma/alpaca-cleaned is a curated version of the original Alpaca Dataset released by Stanford.
6. lmsys/lmsys-chat-1m with 1 million real-world conversations with 25 state-of-the-art LLMs offers diverse use cases, like content moderation, safety benchmarks, and training instruction-following models.
7. allenai/dolma from Allen AI includes 3T tokens from a diverse mix of web content, academic publications, code, books, and encyclopedic materials.
Math datasets:
1. HuggingFaceTB/finemath consists of educational math content and has two versions: 34B tokens and 54B tokens.
2. amphora/QwQ-LongCoT-130K for training O1-like LLMs.
3. openai/gsm8k for training multi-step reasoning.
posted
an
update
about 1 month ago
Post
3073
**15 Agentic Systems and Frameworks of 2024**
This year, we started our “AI Agents and Agentic Workflows” series (https://www.turingpost.com/t/AI-Agents) to explore everything about AI agents step by step: all the vocabulary, how they work, and how to build them.
The huge interest in this series and the large number of studies conducted on agents showed that it was one of the most popular and important themes of the year. In 2025, most likely, agents will reach new highs – we will be covering that for you. Now, let’s review the agentic systems that have emerged this year.
Here is a list of 15 agentic systems and frameworks of 2024:
1. GUI Agents: A Survey (2412.13501)
2. Large Language Models Orchestrating Structured Reasoning Achieve Kaggle Grandmaster Level (2411.03562)
3. The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery (2408.06292)
4. MALT: Improving Reasoning with Multi-Agent LLM Training (2412.01928)
5. Agent S: An Open Agentic Framework that Uses Computers Like a Human (2410.08164)
6. Automated Design of Agentic Systems (2408.08435)
7. AgentInstruct: Toward Generative Teaching with Agentic Flows (2407.03502)
8. AgentStore: Scalable Integration of Heterogeneous Agents As Specialized Generalist Computer Assistant (2410.18603)
9. WALL-E: World Alignment by Rule Learning Improves World Model-based LLM Agents (2410.07484)
10. Generative Agent Simulations of 1,000 People (2411.10109)
11. DynaSaur: Large Language Agents Beyond Predefined Actions (2411.01747)
12. PRefLexOR: Preference-based Recursive Language Modeling for Exploratory Optimization of Reasoning and Agentic Thinking (2410.12375)
13. Generative World Explorer (2411.11844)
14. Bel Esprit: Multi-Agent Framework for Building AI Model Pipelines (2412.14684)
15. AutoKaggle: A Multi-Agent Framework for Autonomous Data Science Competitions (2410.20424)
Thanks for reading Turing Post!
Subscribe to receive new posts straight into your inbox -> https://www.turingpost.com/subscribe
This year, we started our “AI Agents and Agentic Workflows” series (https://www.turingpost.com/t/AI-Agents) to explore everything about AI agents step by step: all the vocabulary, how they work, and how to build them.
The huge interest in this series and the large number of studies conducted on agents showed that it was one of the most popular and important themes of the year. In 2025, most likely, agents will reach new highs – we will be covering that for you. Now, let’s review the agentic systems that have emerged this year.
Here is a list of 15 agentic systems and frameworks of 2024:
1. GUI Agents: A Survey (2412.13501)
2. Large Language Models Orchestrating Structured Reasoning Achieve Kaggle Grandmaster Level (2411.03562)
3. The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery (2408.06292)
4. MALT: Improving Reasoning with Multi-Agent LLM Training (2412.01928)
5. Agent S: An Open Agentic Framework that Uses Computers Like a Human (2410.08164)
6. Automated Design of Agentic Systems (2408.08435)
7. AgentInstruct: Toward Generative Teaching with Agentic Flows (2407.03502)
8. AgentStore: Scalable Integration of Heterogeneous Agents As Specialized Generalist Computer Assistant (2410.18603)
9. WALL-E: World Alignment by Rule Learning Improves World Model-based LLM Agents (2410.07484)
10. Generative Agent Simulations of 1,000 People (2411.10109)
11. DynaSaur: Large Language Agents Beyond Predefined Actions (2411.01747)
12. PRefLexOR: Preference-based Recursive Language Modeling for Exploratory Optimization of Reasoning and Agentic Thinking (2410.12375)
13. Generative World Explorer (2411.11844)
14. Bel Esprit: Multi-Agent Framework for Building AI Model Pipelines (2412.14684)
15. AutoKaggle: A Multi-Agent Framework for Autonomous Data Science Competitions (2410.20424)
Thanks for reading Turing Post!
Subscribe to receive new posts straight into your inbox -> https://www.turingpost.com/subscribe
reacted to
clem's
post with 🔥
about 1 month ago
Post
4583
Six predictions for AI in 2025 (and a review of how my 2024 predictions turned out):
- There will be the first major public protest related to AI
- A big company will see its market cap divided by two or more because of AI
- At least 100,000 personal AI robots will be pre-ordered
- China will start to lead the AI race (as a consequence of leading the open-source AI race).
- There will be big breakthroughs in AI for biology and chemistry.
- We will begin to see the economic and employment growth potential of AI, with 15M AI builders on Hugging Face.
How my predictions for 2024 turned out:
- A hyped AI company will go bankrupt or get acquired for a ridiculously low price
✅ (Inflexion, AdeptAI,...)
- Open-source LLMs will reach the level of the best closed-source LLMs
✅ with QwQ and dozens of others
- Big breakthroughs in AI for video, time-series, biology and chemistry
✅ for video 🔴for time-series, biology and chemistry
- We will talk much more about the cost (monetary and environmental) of AI
✅Monetary 🔴Environmental (😢)
- A popular media will be mostly AI-generated
✅ with NotebookLM by Google
- 10 millions AI builders on Hugging Face leading to no increase of unemployment
🔜currently 7M of AI builders on Hugging Face
- There will be the first major public protest related to AI
- A big company will see its market cap divided by two or more because of AI
- At least 100,000 personal AI robots will be pre-ordered
- China will start to lead the AI race (as a consequence of leading the open-source AI race).
- There will be big breakthroughs in AI for biology and chemistry.
- We will begin to see the economic and employment growth potential of AI, with 15M AI builders on Hugging Face.
How my predictions for 2024 turned out:
- A hyped AI company will go bankrupt or get acquired for a ridiculously low price
✅ (Inflexion, AdeptAI,...)
- Open-source LLMs will reach the level of the best closed-source LLMs
✅ with QwQ and dozens of others
- Big breakthroughs in AI for video, time-series, biology and chemistry
✅ for video 🔴for time-series, biology and chemistry
- We will talk much more about the cost (monetary and environmental) of AI
✅Monetary 🔴Environmental (😢)
- A popular media will be mostly AI-generated
✅ with NotebookLM by Google
- 10 millions AI builders on Hugging Face leading to no increase of unemployment
🔜currently 7M of AI builders on Hugging Face
- 4 replies
replied to
their
post
about 1 month ago
We also got Yoshua Bengio's story: "My own insight really became strong in the context of the machine translation task. Prior to our introduction of attention, we were using a recurrent network that read the whole input source language sequence and then generated the translated target language sequence. However, this is not at all how humans translate. Humans pay very particular attention to just one word or a few input words at a time, in their context, to decide on the next word (or few words) to generate to form the sequence of words in the translation. "
This is the paper he mentions as the one that influenced them: https://www.cs.toronto.edu/~hinton/absps/nips_eyebm.pdf
You can find his full reply here: https://www.turingpost.com/p/attention
reacted to
their
post with ❤️
about 2 months ago
Post
2849
TL;DR: The Story of Attention's Development by
@karpathy
Origin: First proposed in 2014 by @Dzmitry Bahdanau, @KyunghyunCho , and Yoshua Bengio in Neural Machine Translation by Jointly Learning to Align and Translate (1409.0473) . Inspired by cognitive processes and later renamed from "RNNSearch."
Key Idea: A data-dependent weighted average for pooling and communication, enabling flexible and powerful neural network connections.
Breakthrough: Bahdanau's "soft search" mechanism (softmax + weighted averaging) solved encoder-decoder bottlenecks in machine translation.
Transformer Revolution: Attention Is All You Need (1706.03762) (2017) by @ashishvaswanigoogle et al. simplified architectures by stacking attention layers, introducing multi-headed attention and positional encodings.
Legacy: Attention replaced RNNs, driving modern AI systems like ChatGPT. It emerged independently but was influenced by contemporaneous work like Alex Graves’s Neural Turing Machines (1410.5401) and Jason Weston’s Memory Networks (1410.3916) .
Attention to history: Jürgen Schmidhuber claims his 1992 Fast Weight Programmers anticipated modern attention mechanisms. While conceptually similar, the term “attention” was absent, and there’s no evidence it influenced Bahdanau, Cho, and Bengio’s 2014 work. Paying attention (!) to history might have brought us to genAI earlier – but credit for the breakthrough still goes to Montreal.
Referenced Papers:
Attention Origin: Neural Machine Translation by Jointly Learning to Align and Translate (1409.0473)
Transformers: Attention Is All You Need (1706.03762)
Alex Graves' Work: Neural Turing Machines (1410.5401), Generating Sequences With Recurrent Neural Networks (1308.0850)
Jason Weston @spermwhale 's Memory Networks (1410.3916)
Sequence to Sequence Learning with Neural Networks (1409.3215) by Ilya Sutskever ( @ilyasut ), Oriol Vinyals, Quoc V. Le
Who else deserves recognition in this groundbreaking narrative of innovation? Let’s ensure every contributor gets the credit they deserve. Leave a comment below 👇🏻🤗
Origin: First proposed in 2014 by @Dzmitry Bahdanau, @KyunghyunCho , and Yoshua Bengio in Neural Machine Translation by Jointly Learning to Align and Translate (1409.0473) . Inspired by cognitive processes and later renamed from "RNNSearch."
Key Idea: A data-dependent weighted average for pooling and communication, enabling flexible and powerful neural network connections.
Breakthrough: Bahdanau's "soft search" mechanism (softmax + weighted averaging) solved encoder-decoder bottlenecks in machine translation.
Transformer Revolution: Attention Is All You Need (1706.03762) (2017) by @ashishvaswanigoogle et al. simplified architectures by stacking attention layers, introducing multi-headed attention and positional encodings.
Legacy: Attention replaced RNNs, driving modern AI systems like ChatGPT. It emerged independently but was influenced by contemporaneous work like Alex Graves’s Neural Turing Machines (1410.5401) and Jason Weston’s Memory Networks (1410.3916) .
Attention to history: Jürgen Schmidhuber claims his 1992 Fast Weight Programmers anticipated modern attention mechanisms. While conceptually similar, the term “attention” was absent, and there’s no evidence it influenced Bahdanau, Cho, and Bengio’s 2014 work. Paying attention (!) to history might have brought us to genAI earlier – but credit for the breakthrough still goes to Montreal.
Referenced Papers:
Attention Origin: Neural Machine Translation by Jointly Learning to Align and Translate (1409.0473)
Transformers: Attention Is All You Need (1706.03762)
Alex Graves' Work: Neural Turing Machines (1410.5401), Generating Sequences With Recurrent Neural Networks (1308.0850)
Jason Weston @spermwhale 's Memory Networks (1410.3916)
Sequence to Sequence Learning with Neural Networks (1409.3215) by Ilya Sutskever ( @ilyasut ), Oriol Vinyals, Quoc V. Le
Who else deserves recognition in this groundbreaking narrative of innovation? Let’s ensure every contributor gets the credit they deserve. Leave a comment below 👇🏻🤗
-
- 5 replies
replied to
their
post
about 2 months ago
- Jürgen is always right!
- If Jürgen is wrong, see 1.
reacted to
their
post with 🔥
about 2 months ago
Post
2849
TL;DR: The Story of Attention's Development by
@karpathy
Origin: First proposed in 2014 by @Dzmitry Bahdanau, @KyunghyunCho , and Yoshua Bengio in Neural Machine Translation by Jointly Learning to Align and Translate (1409.0473) . Inspired by cognitive processes and later renamed from "RNNSearch."
Key Idea: A data-dependent weighted average for pooling and communication, enabling flexible and powerful neural network connections.
Breakthrough: Bahdanau's "soft search" mechanism (softmax + weighted averaging) solved encoder-decoder bottlenecks in machine translation.
Transformer Revolution: Attention Is All You Need (1706.03762) (2017) by @ashishvaswanigoogle et al. simplified architectures by stacking attention layers, introducing multi-headed attention and positional encodings.
Legacy: Attention replaced RNNs, driving modern AI systems like ChatGPT. It emerged independently but was influenced by contemporaneous work like Alex Graves’s Neural Turing Machines (1410.5401) and Jason Weston’s Memory Networks (1410.3916) .
Attention to history: Jürgen Schmidhuber claims his 1992 Fast Weight Programmers anticipated modern attention mechanisms. While conceptually similar, the term “attention” was absent, and there’s no evidence it influenced Bahdanau, Cho, and Bengio’s 2014 work. Paying attention (!) to history might have brought us to genAI earlier – but credit for the breakthrough still goes to Montreal.
Referenced Papers:
Attention Origin: Neural Machine Translation by Jointly Learning to Align and Translate (1409.0473)
Transformers: Attention Is All You Need (1706.03762)
Alex Graves' Work: Neural Turing Machines (1410.5401), Generating Sequences With Recurrent Neural Networks (1308.0850)
Jason Weston @spermwhale 's Memory Networks (1410.3916)
Sequence to Sequence Learning with Neural Networks (1409.3215) by Ilya Sutskever ( @ilyasut ), Oriol Vinyals, Quoc V. Le
Who else deserves recognition in this groundbreaking narrative of innovation? Let’s ensure every contributor gets the credit they deserve. Leave a comment below 👇🏻🤗
Origin: First proposed in 2014 by @Dzmitry Bahdanau, @KyunghyunCho , and Yoshua Bengio in Neural Machine Translation by Jointly Learning to Align and Translate (1409.0473) . Inspired by cognitive processes and later renamed from "RNNSearch."
Key Idea: A data-dependent weighted average for pooling and communication, enabling flexible and powerful neural network connections.
Breakthrough: Bahdanau's "soft search" mechanism (softmax + weighted averaging) solved encoder-decoder bottlenecks in machine translation.
Transformer Revolution: Attention Is All You Need (1706.03762) (2017) by @ashishvaswanigoogle et al. simplified architectures by stacking attention layers, introducing multi-headed attention and positional encodings.
Legacy: Attention replaced RNNs, driving modern AI systems like ChatGPT. It emerged independently but was influenced by contemporaneous work like Alex Graves’s Neural Turing Machines (1410.5401) and Jason Weston’s Memory Networks (1410.3916) .
Attention to history: Jürgen Schmidhuber claims his 1992 Fast Weight Programmers anticipated modern attention mechanisms. While conceptually similar, the term “attention” was absent, and there’s no evidence it influenced Bahdanau, Cho, and Bengio’s 2014 work. Paying attention (!) to history might have brought us to genAI earlier – but credit for the breakthrough still goes to Montreal.
Referenced Papers:
Attention Origin: Neural Machine Translation by Jointly Learning to Align and Translate (1409.0473)
Transformers: Attention Is All You Need (1706.03762)
Alex Graves' Work: Neural Turing Machines (1410.5401), Generating Sequences With Recurrent Neural Networks (1308.0850)
Jason Weston @spermwhale 's Memory Networks (1410.3916)
Sequence to Sequence Learning with Neural Networks (1409.3215) by Ilya Sutskever ( @ilyasut ), Oriol Vinyals, Quoc V. Le
Who else deserves recognition in this groundbreaking narrative of innovation? Let’s ensure every contributor gets the credit they deserve. Leave a comment below 👇🏻🤗
-
-
-
- 5 replies
replied to
their
post
about 2 months ago
The original post by Karpathy https://x.com/karpathy/status/1864023344435380613