diff --git "a/benchmarking.ipynb" "b/benchmarking.ipynb" --- "a/benchmarking.ipynb" +++ "b/benchmarking.ipynb" @@ -8,33 +8,23 @@ "\n", "* After exploring a variety of possible benchmarks, I decided to focus on SemScore, which is a semantic similarity metric. \n", "* The idea is to evaluate how well the agent can answer questions that are syntactically and semantically similar to the reference answers.\n", - "* One challenge is that, because I chose to " + "* It uses cosine similarity of embedding vectors to measure the semantic similarity between the predicted answer and the reference answer.\n", + "* It is a good metric for evaluating the quality of the agent's answers, but it does not take into account the existence of multiple acceptable answers.\n", + "* It also does not take into account the quality of the question, which is as important as the quality of the answer." ] }, { "cell_type": "markdown", "metadata": {}, - "source": [] + "source": [ + "### Setup" + ] }, { "cell_type": "code", - "execution_count": 1, + "execution_count": 211, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Initializing Data...\n", - "Download: True\n", - "Loading data...\n", - "Raw Data loaded\n", - "Chroma DB already exists\n", - "Loading index...\n", - "Index loaded\n" - ] - } - ], + "outputs": [], "source": [ "import os\n", "import numpy as np\n", @@ -46,327 +36,272 @@ "from semscore import EmbeddingModelWrapper\n", "from statistics import mean\n", "from agent import get_agent\n", - "from dotenv import load_dotenv\n", + "from openai import OpenAI\n", "from prompts import SUCCINCT_SQUAD_REACT_CODE_SYSTEM_PROMPT, FOCUSED_SQUAD_REACT_CODE_SYSTEM_PROMPT\n", - "\n", + "import re\n", + "from string import punctuation\n", + "from nltk.corpus import stopwords\n", + "import matplotlib.pyplot as plt\n", + "import seaborn as sns\n", + "from dotenv import load_dotenv\n", "load_dotenv() # Load OPENAI_API_KEY from .env (not included in repo)\n", "\n", "SAMPLES_DIR = \"samples\"\n", + "BENCHMARKS_DIR = \"benchmarks\"\n", + "STOP_WORDS = set(stopwords.words('english'))\n", "\n", "def display_text_df(df):\n", " display(df.style.set_properties(**{'white-space': 'pre-wrap'}).set_table_styles(\n", " [{'selector': 'th', 'props': [('text-align', 'left')]},\n", " {'selector': 'td', 'props': [('text-align', 'left')]}\n", " ]\n", - " ).hide())\n" - ] - }, - { - "cell_type": "code", - "execution_count": 2, - "metadata": {}, - "outputs": [], - "source": [ - "from data import get_data\n", - "data = get_data(download=False)\n" + " ))\n" ] }, { - "cell_type": "code", - "execution_count": 3, + "cell_type": "markdown", "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "(87599, 4)" - ] - }, - "execution_count": 3, - "metadata": {}, - "output_type": "execute_result" - } - ], "source": [ - "data.df.shape" + "#### Load the data" ] }, { "cell_type": "code", - "execution_count": 4, + "execution_count": 199, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", - "\n", + "
\n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", "
TitleContextQuestionAnswerTitleContextQuestionAnswer
University_of_Notre_DameArchitecturally, the school has a Catholic character. Atop the Main Building's gold dome is a golden statue of the Virgin Mary. Immediately in front of the Main Building and facing it, is a copper statue of Christ with arms upraised with the legend \"Venite Ad Me Omnes\". Next to the Main Building is the Basilica of the Sacred Heart. Immediately behind the basilica is the Grotto, a Marian place of prayer and reflection. It is a replica of the grotto at Lourdes, France where the Virgin Mary reputedly appeared to Saint Bernadette Soubirous in 1858. At the end of the main drive (and in a direct line that connects through 3 statues and the Gold Dome), is a simple, modern stone statue of Mary.To whom did the Virgin Mary allegedly appear in 1858 in Lourdes France?Saint Bernadette SoubirousUniversity_of_Notre_DameArchitecturally, the school has a Catholic character. Atop the Main Building's gold dome is a golden statue of the Virgin Mary. Immediately in front of the Main Building and facing it, is a copper statue of Christ with arms upraised with the legend \"Venite Ad Me Omnes\". Next to the Main Building is the Basilica of the Sacred Heart. Immediately behind the basilica is the Grotto, a Marian place of prayer and reflection. It is a replica of the grotto at Lourdes, France where the Virgin Mary reputedly appeared to Saint Bernadette Soubirous in 1858. At the end of the main drive (and in a direct line that connects through 3 statues and the Gold Dome), is a simple, modern stone statue of Mary.To whom did the Virgin Mary allegedly appear in 1858 in Lourdes France?Saint Bernadette Soubirous
University_of_Notre_DameArchitecturally, the school has a Catholic character. Atop the Main Building's gold dome is a golden statue of the Virgin Mary. Immediately in front of the Main Building and facing it, is a copper statue of Christ with arms upraised with the legend \"Venite Ad Me Omnes\". Next to the Main Building is the Basilica of the Sacred Heart. Immediately behind the basilica is the Grotto, a Marian place of prayer and reflection. It is a replica of the grotto at Lourdes, France where the Virgin Mary reputedly appeared to Saint Bernadette Soubirous in 1858. At the end of the main drive (and in a direct line that connects through 3 statues and the Gold Dome), is a simple, modern stone statue of Mary.What is in front of the Notre Dame Main Building?a copper statue of ChristUniversity_of_Notre_DameArchitecturally, the school has a Catholic character. Atop the Main Building's gold dome is a golden statue of the Virgin Mary. Immediately in front of the Main Building and facing it, is a copper statue of Christ with arms upraised with the legend \"Venite Ad Me Omnes\". Next to the Main Building is the Basilica of the Sacred Heart. Immediately behind the basilica is the Grotto, a Marian place of prayer and reflection. It is a replica of the grotto at Lourdes, France where the Virgin Mary reputedly appeared to Saint Bernadette Soubirous in 1858. At the end of the main drive (and in a direct line that connects through 3 statues and the Gold Dome), is a simple, modern stone statue of Mary.What is in front of the Notre Dame Main Building?a copper statue of Christ
University_of_Notre_DameArchitecturally, the school has a Catholic character. Atop the Main Building's gold dome is a golden statue of the Virgin Mary. Immediately in front of the Main Building and facing it, is a copper statue of Christ with arms upraised with the legend \"Venite Ad Me Omnes\". Next to the Main Building is the Basilica of the Sacred Heart. Immediately behind the basilica is the Grotto, a Marian place of prayer and reflection. It is a replica of the grotto at Lourdes, France where the Virgin Mary reputedly appeared to Saint Bernadette Soubirous in 1858. At the end of the main drive (and in a direct line that connects through 3 statues and the Gold Dome), is a simple, modern stone statue of Mary.The Basilica of the Sacred heart at Notre Dame is beside to which structure?the Main BuildingUniversity_of_Notre_DameArchitecturally, the school has a Catholic character. Atop the Main Building's gold dome is a golden statue of the Virgin Mary. Immediately in front of the Main Building and facing it, is a copper statue of Christ with arms upraised with the legend \"Venite Ad Me Omnes\". Next to the Main Building is the Basilica of the Sacred Heart. Immediately behind the basilica is the Grotto, a Marian place of prayer and reflection. It is a replica of the grotto at Lourdes, France where the Virgin Mary reputedly appeared to Saint Bernadette Soubirous in 1858. At the end of the main drive (and in a direct line that connects through 3 statues and the Gold Dome), is a simple, modern stone statue of Mary.The Basilica of the Sacred heart at Notre Dame is beside to which structure?the Main Building
\n" ], "text/plain": [ - "" + "" ] }, "metadata": {}, "output_type": "display_data" + }, + { + "data": { + "text/plain": [ + "(87599, 4)" + ] + }, + "execution_count": 199, + "metadata": {}, + "output_type": "execute_result" } ], "source": [ - "display_text_df(data.df.head(3))\n" + "from data import get_data\n", + "data = get_data(download=False)\n", + "display_text_df(data.df.head(3))\n", + "data.df.shape" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Sample 100 random rows from the data" ] }, { "cell_type": "code", - "execution_count": 17, + "execution_count": 200, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", - "\n", + "
\n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", "
TitleContextQuestionAnswerTitleContextQuestionAnswer
Institute_of_technologyThe world's first institution of technology or technical university with tertiary technical education is the Banská Akadémia in Banská Štiavnica, Slovakia, founded in 1735, Academy since December 13, 1762 established by queen Maria Theresa in order to train specialists of silver and gold mining and metallurgy in neighbourhood. Teaching started in 1764. Later the department of Mathematics, Mechanics and Hydraulics and department of Forestry were settled. University buildings are still at their place today and are used for teaching. University has launched the first book of electrotechnics in the world.What year was the Banská Akadémia founded?1735Institute_of_technologyThe world's first institution of technology or technical university with tertiary technical education is the Banská Akadémia in Banská Štiavnica, Slovakia, founded in 1735, Academy since December 13, 1762 established by queen Maria Theresa in order to train specialists of silver and gold mining and metallurgy in neighbourhood. Teaching started in 1764. Later the department of Mathematics, Mechanics and Hydraulics and department of Forestry were settled. University buildings are still at their place today and are used for teaching. University has launched the first book of electrotechnics in the world.What year was the Banská Akadémia founded?1735
Film_speedThe standard specifies how speed ratings should be reported by the camera. If the noise-based speed (40:1) is higher than the saturation-based speed, the noise-based speed should be reported, rounded downwards to a standard value (e.g. 200, 250, 320, or 400). The rationale is that exposure according to the lower saturation-based speed would not result in a visibly better image. In addition, an exposure latitude can be specified, ranging from the saturation-based speed to the 10:1 noise-based speed. If the noise-based speed (40:1) is lower than the saturation-based speed, or undefined because of high noise, the saturation-based speed is specified, rounded upwards to a standard value, because using the noise-based speed would lead to overexposed images. The camera may also report the SOS-based speed (explicitly as being an SOS speed), rounded to the nearest standard speed rating.What is another speed that can also be reported by the camera?SOS-based speedFilm_speedThe standard specifies how speed ratings should be reported by the camera. If the noise-based speed (40:1) is higher than the saturation-based speed, the noise-based speed should be reported, rounded downwards to a standard value (e.g. 200, 250, 320, or 400). The rationale is that exposure according to the lower saturation-based speed would not result in a visibly better image. In addition, an exposure latitude can be specified, ranging from the saturation-based speed to the 10:1 noise-based speed. If the noise-based speed (40:1) is lower than the saturation-based speed, or undefined because of high noise, the saturation-based speed is specified, rounded upwards to a standard value, because using the noise-based speed would lead to overexposed images. The camera may also report the SOS-based speed (explicitly as being an SOS speed), rounded to the nearest standard speed rating.What is another speed that can also be reported by the camera?SOS-based speed
SumerThe most impressive and famous of Sumerian buildings are the ziggurats, large layered platforms which supported temples. Sumerian cylinder seals also depict houses built from reeds not unlike those built by the Marsh Arabs of Southern Iraq until as recently as 400 CE. The Sumerians also developed the arch, which enabled them to develop a strong type of dome. They built this by constructing and linking several arches. Sumerian temples and palaces made use of more advanced materials and techniques,[citation needed] such as buttresses, recesses, half columns, and clay nails.Where were the use of advanced materials and techniques on display in Sumer?Sumerian temples and palacesSumerThe most impressive and famous of Sumerian buildings are the ziggurats, large layered platforms which supported temples. Sumerian cylinder seals also depict houses built from reeds not unlike those built by the Marsh Arabs of Southern Iraq until as recently as 400 CE. The Sumerians also developed the arch, which enabled them to develop a strong type of dome. They built this by constructing and linking several arches. Sumerian temples and palaces made use of more advanced materials and techniques,[citation needed] such as buttresses, recesses, half columns, and clay nails.Where were the use of advanced materials and techniques on display in Sumer?Sumerian temples and palaces
Ann_Arbor,_MichiganAnn Arbor has a council-manager form of government. The City Council has 11 voting members: the mayor and 10 city council members. The mayor and city council members serve two-year terms: the mayor is elected every even-numbered year, while half of the city council members are up for election annually (five in even-numbered and five in odd-numbered years). Two council members are elected from each of the city's five wards. The mayor is elected citywide. The mayor is the presiding officer of the City Council and has the power to appoint all Council committee members as well as board and commission members, with the approval of the City Council. The current mayor of Ann Arbor is Christopher Taylor, a Democrat who was elected as mayor in 2014. Day-to-day city operations are managed by a city administrator chosen by the city council.Who is elected every even numbered year?mayorAnn_Arbor,_MichiganAnn Arbor has a council-manager form of government. The City Council has 11 voting members: the mayor and 10 city council members. The mayor and city council members serve two-year terms: the mayor is elected every even-numbered year, while half of the city council members are up for election annually (five in even-numbered and five in odd-numbered years). Two council members are elected from each of the city's five wards. The mayor is elected citywide. The mayor is the presiding officer of the City Council and has the power to appoint all Council committee members as well as board and commission members, with the approval of the City Council. The current mayor of Ann Arbor is Christopher Taylor, a Democrat who was elected as mayor in 2014. Day-to-day city operations are managed by a city administrator chosen by the city council.Who is elected every even numbered year?mayor
John_von_NeumannShortly before his death, when he was already quite ill, von Neumann headed the United States government's top secret ICBM committee, and it would sometimes meet in his home. Its purpose was to decide on the feasibility of building an ICBM large enough to carry a thermonuclear weapon. Von Neumann had long argued that while the technical obstacles were sizable, they could be overcome in time. The SM-65 Atlas passed its first fully functional test in 1959, two years after his death. The feasibility of an ICBM owed as much to improved, smaller warheads as it did to developments in rocketry, and his understanding of the former made his advice invaluable.What was the purpose of top secret ICBM committee?decide on the feasibility of building an ICBM large enough to carry a thermonuclear weaponJohn_von_NeumannShortly before his death, when he was already quite ill, von Neumann headed the United States government's top secret ICBM committee, and it would sometimes meet in his home. Its purpose was to decide on the feasibility of building an ICBM large enough to carry a thermonuclear weapon. Von Neumann had long argued that while the technical obstacles were sizable, they could be overcome in time. The SM-65 Atlas passed its first fully functional test in 1959, two years after his death. The feasibility of an ICBM owed as much to improved, smaller warheads as it did to developments in rocketry, and his understanding of the former made his advice invaluable.What was the purpose of top secret ICBM committee?decide on the feasibility of building an ICBM large enough to carry a thermonuclear weapon
\n" ], "text/plain": [ - "" + "" ] }, "metadata": {}, "output_type": "display_data" + }, + { + "data": { + "text/plain": [ + "(100, 4)" + ] + }, + "execution_count": 200, + "metadata": {}, + "output_type": "execute_result" } ], "source": [ "np.random.seed(42)\n", "# Select 10 random rows from data.df\n", - "dfSample = data.df.sample(n=300)\n", - "display_text_df(dfSample.head())" - ] - }, - { - "cell_type": "code", - "execution_count": 18, - "metadata": {}, - "outputs": [], - "source": [ - "# use local gpt to synthesize questions with context\n", - "\n", - "synth_system_prompt = \"\"\"\n", - "You are an expert at clarifying what questions are really asking for.\n", - "\n", - "You will be given a question, a title and context.\n", - "Your task is come up with a new version of the question that resolves ambiguities by adding only and exactly the necessary details from the title and context in a way that clarifies the question without changing the meaning or intent of the question.\n", - "\n", - "For example: \n", - "Question: Who does M fight with?\n", - "Title: Spectre_(2015_film)\n", - "Context: Bond and Swann return to London where they meet M, Bill Tanner, Q, and Moneypenny; they intend to arrest C and stop Nine Eyes from going online. Swann leaves Bond, telling him she cannot be part of a life involving espionage, and is subsequently kidnapped. On the way, the group is ambushed and Bond is kidnapped, but the rest still proceed with the plan. After Q succeeds in preventing the Nine Eyes from going online, a brief struggle between M and C ends with the latter falling to his death. Meanwhile, Bond is taken to the old MI6 building, which is scheduled for demolition, and frees himself. Moving throughout the ruined labyrinth, he encounters a disfigured Blofeld, who tells him that he has three minutes to escape the building before explosives are detonated or die trying to save Swann. Bond finds Swann and the two escape by boat as the building collapses. Bond shoots down Blofeld's helicopter, which crashes onto Westminster Bridge. As Blofeld crawls away from the wreckage, Bond confronts him but ultimately leaves him to be arrested by M. Bond leaves the bridge with Swann.\n", - "Response: Who does M struggle with during the events of Spectre (2015)?\n", - "\"\"\"\n", - "\n", - "synth_user_prompt = \"\"\"\n", - "Question: {question}\n", - "Title: {title}\n", - "Context: {context}\n", - "\"\"\"\n" + "dfSample = data.df.sample(n=100)\n", + "display_text_df(dfSample.head())\n", + "dfSample.shape" ] }, { - "cell_type": "code", - "execution_count": null, + "cell_type": "markdown", "metadata": {}, - "outputs": [], "source": [ - "# Example: reuse your existing OpenAI setup\n", - "from openai import OpenAI\n", + "## Synthesize Unambiguous Questions\n", "\n", - "# Point to the local server\n", - "# client = OpenAI(base_url=\"http://localhost:1234/v1\", api_key=\"lm-studio\")\n", - "client = OpenAI()\n", - "\n", - "synth_answers = []\n", - "for title, context, question, answer in tqdm(dfSample.values):\n", - " completion = client.chat.completions.create(\n", - " # model=\"lmstudio-community/Meta-Llama-3.1-8B-Instruct-GGUF\",\n", - " model=\"gpt-4o-mini-2024-07-18\",\n", - " messages=[\n", - " {\"role\": \"system\", \"content\": synth_system_prompt},\n", - " {\"role\": \"user\", \"content\": synth_user_prompt.format(question=question, title=title, context=context)}\n", - " ],\n", - " temperature=0.7,\n", - " )\n", - " synth_answers.append(completion.choices[0].message.content)\n", - "\n", - "dfSample[\"Synthesized Question\"] = synth_answers" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "os.makedirs(SAMPLES_DIR, exist_ok=True)\n", - "dfSample.to_pickle(os.path.join(SAMPLES_DIR, f\"samples.pkl\")) \n" + "* Because the solution is Closed Generative QA, the raw questions in the dataset may result in unreasonable standards in the benchmark due to their ambiguity.\n", + "* Therefore, we need to synthesize unambiguous questions.\n", + "* For this, we will use GPT-4o-mini and a simple prompt, one-shot prompt to synthesize the questions." ] }, { "cell_type": "code", - "execution_count": 48, + "execution_count": 201, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", - "\n", + "
\n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", "
TitleContextQuestionAnswerSynthesized QuestionTitleContextQuestionAnswerSynthesized Question
Institute_of_technologyThe world's first institution of technology or technical university with tertiary technical education is the Banská Akadémia in Banská Štiavnica, Slovakia, founded in 1735, Academy since December 13, 1762 established by queen Maria Theresa in order to train specialists of silver and gold mining and metallurgy in neighbourhood. Teaching started in 1764. Later the department of Mathematics, Mechanics and Hydraulics and department of Forestry were settled. University buildings are still at their place today and are used for teaching. University has launched the first book of electrotechnics in the world.What year was the Banská Akadémia founded?1735What year was the Banská Akadémia, the world's first institution of technology, founded in Banská Štiavnica, Slovakia?
Film_speedThe standard specifies how speed ratings should be reported by the camera. If the noise-based speed (40:1) is higher than the saturation-based speed, the noise-based speed should be reported, rounded downwards to a standard value (e.g. 200, 250, 320, or 400). The rationale is that exposure according to the lower saturation-based speed would not result in a visibly better image. In addition, an exposure latitude can be specified, ranging from the saturation-based speed to the 10:1 noise-based speed. If the noise-based speed (40:1) is lower than the saturation-based speed, or undefined because of high noise, the saturation-based speed is specified, rounded upwards to a standard value, because using the noise-based speed would lead to overexposed images. The camera may also report the SOS-based speed (explicitly as being an SOS speed), rounded to the nearest standard speed rating.What is another speed that can also be reported by the camera?SOS-based speedWhat is another speed rating that can also be reported by the camera in addition to the noise-based and saturation-based speeds?
SumerThe most impressive and famous of Sumerian buildings are the ziggurats, large layered platforms which supported temples. Sumerian cylinder seals also depict houses built from reeds not unlike those built by the Marsh Arabs of Southern Iraq until as recently as 400 CE. The Sumerians also developed the arch, which enabled them to develop a strong type of dome. They built this by constructing and linking several arches. Sumerian temples and palaces made use of more advanced materials and techniques,[citation needed] such as buttresses, recesses, half columns, and clay nails.Where were the use of advanced materials and techniques on display in Sumer?Sumerian temples and palacesWhere were the advanced materials and techniques, such as buttresses and arches, used in Sumerian temples and palaces on display?Institute_of_technologyThe world's first institution of technology or technical university with tertiary technical education is the Banská Akadémia in Banská Štiavnica, Slovakia, founded in 1735, Academy since December 13, 1762 established by queen Maria Theresa in order to train specialists of silver and gold mining and metallurgy in neighbourhood. Teaching started in 1764. Later the department of Mathematics, Mechanics and Hydraulics and department of Forestry were settled. University buildings are still at their place today and are used for teaching. University has launched the first book of electrotechnics in the world.What year was the Banská Akadémia founded?1735What year was the Banská Akadémia, the world's first institution of technology, founded in Banská Štiavnica, Slovakia?
Ann_Arbor,_MichiganAnn Arbor has a council-manager form of government. The City Council has 11 voting members: the mayor and 10 city council members. The mayor and city council members serve two-year terms: the mayor is elected every even-numbered year, while half of the city council members are up for election annually (five in even-numbered and five in odd-numbered years). Two council members are elected from each of the city's five wards. The mayor is elected citywide. The mayor is the presiding officer of the City Council and has the power to appoint all Council committee members as well as board and commission members, with the approval of the City Council. The current mayor of Ann Arbor is Christopher Taylor, a Democrat who was elected as mayor in 2014. Day-to-day city operations are managed by a city administrator chosen by the city council.Who is elected every even numbered year?mayorWho is elected as mayor every even-numbered year in Ann Arbor, Michigan?Film_speedThe standard specifies how speed ratings should be reported by the camera. If the noise-based speed (40:1) is higher than the saturation-based speed, the noise-based speed should be reported, rounded downwards to a standard value (e.g. 200, 250, 320, or 400). The rationale is that exposure according to the lower saturation-based speed would not result in a visibly better image. In addition, an exposure latitude can be specified, ranging from the saturation-based speed to the 10:1 noise-based speed. If the noise-based speed (40:1) is lower than the saturation-based speed, or undefined because of high noise, the saturation-based speed is specified, rounded upwards to a standard value, because using the noise-based speed would lead to overexposed images. The camera may also report the SOS-based speed (explicitly as being an SOS speed), rounded to the nearest standard speed rating.What is another speed that can also be reported by the camera?SOS-based speedWhat is another speed rating that can also be reported by the camera in addition to the noise-based and saturation-based speeds?
John_von_NeumannShortly before his death, when he was already quite ill, von Neumann headed the United States government's top secret ICBM committee, and it would sometimes meet in his home. Its purpose was to decide on the feasibility of building an ICBM large enough to carry a thermonuclear weapon. Von Neumann had long argued that while the technical obstacles were sizable, they could be overcome in time. The SM-65 Atlas passed its first fully functional test in 1959, two years after his death. The feasibility of an ICBM owed as much to improved, smaller warheads as it did to developments in rocketry, and his understanding of the former made his advice invaluable.What was the purpose of top secret ICBM committee?decide on the feasibility of building an ICBM large enough to carry a thermonuclear weaponWhat was the purpose of the top secret ICBM committee headed by John von Neumann shortly before his death?SumerThe most impressive and famous of Sumerian buildings are the ziggurats, large layered platforms which supported temples. Sumerian cylinder seals also depict houses built from reeds not unlike those built by the Marsh Arabs of Southern Iraq until as recently as 400 CE. The Sumerians also developed the arch, which enabled them to develop a strong type of dome. They built this by constructing and linking several arches. Sumerian temples and palaces made use of more advanced materials and techniques,[citation needed] such as buttresses, recesses, half columns, and clay nails.Where were the use of advanced materials and techniques on display in Sumer?Sumerian temples and palacesWhere were the advanced materials and techniques, such as buttresses and arches, used in Sumerian temples and palaces on display?
\n" ], "text/plain": [ - "" + "" ] }, "metadata": {}, @@ -378,20 +313,75 @@ "(100, 5)" ] }, - "execution_count": 48, + "execution_count": 201, "metadata": {}, "output_type": "execute_result" } ], "source": [ - "dfSample = pd.read_pickle(os.path.join(SAMPLES_DIR, f\"samples.pkl\"))\n", - "display_text_df(dfSample.head())\n", + "# use local gpt to synthesize questions with context\n", + "\n", + "synth_system_prompt = \"\"\"\n", + "You are an expert at clarifying what questions are really asking for.\n", + "\n", + "You will be given a question, a title and context.\n", + "Your task is come up with a new version of the question that resolves ambiguities \n", + "by adding only and exactly the necessary details from the title and context \n", + "in a way that clarifies the question without changing the meaning or intent of the question.\n", + "\n", + "For example: \n", + "Question: Who does M fight with?\n", + "Title: Spectre_(2015_film)\n", + "Context: Bond and Swann return to London where they meet M, Bill Tanner, Q, and Moneypenny; they intend to arrest C and stop Nine Eyes from going online. Swann leaves Bond, telling him she cannot be part of a life involving espionage, and is subsequently kidnapped. On the way, the group is ambushed and Bond is kidnapped, but the rest still proceed with the plan. After Q succeeds in preventing the Nine Eyes from going online, a brief struggle between M and C ends with the latter falling to his death. Meanwhile, Bond is taken to the old MI6 building, which is scheduled for demolition, and frees himself. Moving throughout the ruined labyrinth, he encounters a disfigured Blofeld, who tells him that he has three minutes to escape the building before explosives are detonated or die trying to save Swann. Bond finds Swann and the two escape by boat as the building collapses. Bond shoots down Blofeld's helicopter, which crashes onto Westminster Bridge. As Blofeld crawls away from the wreckage, Bond confronts him but ultimately leaves him to be arrested by M. Bond leaves the bridge with Swann.\n", + "Response: Who does M struggle with during the events of Spectre (2015)?\n", + "\"\"\"\n", + "\n", + "synth_user_prompt = \"\"\"\n", + "Question: {question}\n", + "Title: {title}\n", + "Context: {context}\n", + "\"\"\"\n", + "\n", + "client = OpenAI()\n", + "\n", + "# if the samples file does not exist, synthesize the questions and save them\n", + "if not os.path.exists(os.path.join(SAMPLES_DIR, f\"samples.pkl\")):\n", + " synth_answers = []\n", + " for title, context, question, answer in tqdm(dfSample.values):\n", + " completion = client.chat.completions.create(\n", + " model=\"gpt-4o-mini-2024-07-18\",\n", + " messages=[\n", + " {\"role\": \"system\", \"content\": synth_system_prompt},\n", + " {\"role\": \"user\", \"content\": synth_user_prompt.format(question=question, title=title, context=context)}\n", + " ],\n", + " temperature=0.7,\n", + " )\n", + " synth_answers.append(completion.choices[0].message.content)\n", + "\n", + " dfSample[\"Synthesized Question\"] = synth_answers\n", + "\n", + " os.makedirs(SAMPLES_DIR, exist_ok=True)\n", + " dfSample.to_pickle(os.path.join(SAMPLES_DIR, f\"samples.pkl\")) \n", + "else:\n", + " # if the samples file exists, load it\n", + " dfSample = pd.read_pickle(os.path.join(SAMPLES_DIR, f\"samples.pkl\"))\n", + "\n", + "display_text_df(dfSample.head(3))\n", "dfSample.shape" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Benchmark the agent\n", + "\n", + "* First, let's test the agent on a single question to see how it performs, show its logs, and the final answer." + ] + }, { "cell_type": "code", - "execution_count": 9, + "execution_count": 202, "metadata": {}, "outputs": [ { @@ -401,9 +391,9 @@ "\u001b[32;20;1m======== New task ========\u001b[0m\n", "\u001b[37;1mWhat year was the Banská Akadémia, the world's first institution of technology, founded in Banská Štiavnica, Slovakia?\u001b[0m\n", "\u001b[33;1m=== Agent thoughts:\u001b[0m\n", - "\u001b[0mThought: I will use the squad_retriever tool to find the founding year of the Banská Akadémia, which is known as the world's first institution of technology located in Banská Štiavnica, Slovakia. I will craft a detailed query to ensure I retrieve the most relevant information.\u001b[0m\n", + "\u001b[0mThought: I will use the squad_retriever tool to find information about the Banská Akadémia, specifically its founding year. I will phrase my query to include details about its significance as the world's first institution of technology located in Banská Štiavnica, Slovakia.\u001b[0m\n", "\u001b[33;1m>>> Agent is executing the code below:\u001b[0m\n", - "\u001b[0m\u001b[38;5;7manswer\u001b[39m\u001b[38;5;7m \u001b[39m\u001b[38;5;109;01m=\u001b[39;00m\u001b[38;5;7m \u001b[39m\u001b[38;5;7msquad_retriever\u001b[39m\u001b[38;5;7m(\u001b[39m\u001b[38;5;7mquery\u001b[39m\u001b[38;5;109;01m=\u001b[39;00m\u001b[38;5;144m\"\u001b[39m\u001b[38;5;144mWhat year was the Banská Akadémia, the world\u001b[39m\u001b[38;5;144m'\u001b[39m\u001b[38;5;144ms first institution of technology, founded in Banská Štiavnica, Slovakia?\u001b[39m\u001b[38;5;144m\"\u001b[39m\u001b[38;5;7m)\u001b[39m\n", + "\u001b[0m\u001b[38;5;7manswer\u001b[39m\u001b[38;5;7m \u001b[39m\u001b[38;5;109;01m=\u001b[39;00m\u001b[38;5;7m \u001b[39m\u001b[38;5;7msquad_retriever\u001b[39m\u001b[38;5;7m(\u001b[39m\u001b[38;5;7mquery\u001b[39m\u001b[38;5;109;01m=\u001b[39;00m\u001b[38;5;144m\"\u001b[39m\u001b[38;5;144mWhat year was the Banská Akadémia founded in Banská Štiavnica, Slovakia, known as the world\u001b[39m\u001b[38;5;144m'\u001b[39m\u001b[38;5;144ms first institution of technology?\u001b[39m\u001b[38;5;144m\"\u001b[39m\u001b[38;5;7m)\u001b[39m\n", "\u001b[38;5;109mprint\u001b[39m\u001b[38;5;7m(\u001b[39m\u001b[38;5;7manswer\u001b[39m\u001b[38;5;7m)\u001b[39m\u001b[0m\n", "\u001b[33;1m====\u001b[0m\n", "\u001b[33;1mPrint outputs:\u001b[0m\n", @@ -413,35 +403,35 @@ "Question: What year was the Banská Akadémia founded?\n", "Acceptable Answers:\n", "['1. 1735']\n", - "Score: 0.8756514910949876\n", + "Score: 0.8805255214632872\n", "===Document===\n", "Title: Institute_of_technology\n", "Context: The world's first institution of technology or technical university with tertiary technical education is the Banská Akadémia in Banská Štiavnica, Slovakia, founded in 1735, Academy since December 13, 1762 established by queen Maria Theresa in order to train specialists of silver and gold mining and metallurgy in neighbourhood. Teaching started in 1764. Later the department of Mathematics, Mechanics and Hydraulics and department of Forestry were settled. University buildings are still at their place today and are used for teaching. University has launched the first book of electrotechnics in the world.\n", "Question: What year did teaching start at the Banská Akadémia?\n", "Acceptable Answers:\n", "['1. 1764']\n", - "Score: 0.8678412986560574\n", + "Score: 0.8732076610524725\n", "\u001b[0m\n", "\u001b[33;1m=== Agent thoughts:\u001b[0m\n", "\u001b[0mThought: From the information retrieved, I learned that the Banská Akadémia was founded in the year 1735. I will now use this information to provide the final answer.\u001b[0m\n", "\u001b[33;1m>>> Agent is executing the code below:\u001b[0m\n", - "\u001b[0m\u001b[38;5;7mfinal_answer\u001b[39m\u001b[38;5;7m(\u001b[39m\u001b[38;5;144m\"\u001b[39m\u001b[38;5;144mThe Banská Akadémia, the world\u001b[39m\u001b[38;5;144m'\u001b[39m\u001b[38;5;144ms first institution of technology, was founded in 1735.\u001b[39m\u001b[38;5;144m\"\u001b[39m\u001b[38;5;7m)\u001b[39m\u001b[0m\n", + "\u001b[0m\u001b[38;5;7mfinal_answer\u001b[39m\u001b[38;5;7m(\u001b[39m\u001b[38;5;144m\"\u001b[39m\u001b[38;5;144mThe Banská Akadémia, the world\u001b[39m\u001b[38;5;144m'\u001b[39m\u001b[38;5;144ms first institution of technology, was founded in the year 1735.\u001b[39m\u001b[38;5;144m\"\u001b[39m\u001b[38;5;7m)\u001b[39m\u001b[0m\n", "\u001b[33;1m====\u001b[0m\n", "\u001b[33;1mPrint outputs:\u001b[0m\n", "\u001b[32;20m\u001b[0m\n", "\u001b[33;1mLast output from code snippet:\u001b[0m\n", - "\u001b[32;20mThe Banská Akadémia, the world's first institution of technology, was founded in 1735.\u001b[0m\n", + "\u001b[32;20mThe Banská Akadémia, the world's first institution of technology, was founded in the year 1735.\u001b[0m\n", "\u001b[32;20;1mFinal answer:\u001b[0m\n", - "\u001b[32;20mThe Banská Akadémia, the world's first institution of technology, was founded in 1735.\u001b[0m\n" + "\u001b[32;20mThe Banská Akadémia, the world's first institution of technology, was founded in the year 1735.\u001b[0m\n" ] }, { "data": { "text/plain": [ - "\"The Banská Akadémia, the world's first institution of technology, was founded in 1735.\"" + "\"The Banská Akadémia, the world's first institution of technology, was founded in the year 1735.\"" ] }, - "execution_count": 9, + "execution_count": 202, "metadata": {}, "output_type": "execute_result" } @@ -456,38 +446,25 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "### Create the agent to be evaluated" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "### Run the agent on the random sample of questions\n", - "\n", - "* Unlike the default Retrieval QA or Open Generative QA of SQuAD, in our use case, the agent would normally be given context in the course of a natural conversation, as the user elaborates on what they want to know. \n", - "* Therefore, for benchmarking, we will provide the context to answer the question in the prompt.\n", + "### Define the benchmark\n", "\n", - "### Use semantic similarity to evaluate the agent's answers against the reference answers\n", + "* We are using semantic similarity to evaluate the agent's answers against the reference answers.\n", + "* During test runs, it became clear that the agent was being penalized for punctuation, stop words, and minor differences in case.\n", + "* Therefore, we will clean the text of the expected and predicted answers before calculating the semantic similarity.\n", "\n", + "### BenchmarkDesign Notes\n", "* One flaw of this approach is that it does not take into account the existence of multiple acceptable answers.\n", - "* Another flaw is that the agent me be unfairly penalized for elaborating on the answer, while this benchmark focuses on only and exactly the one canonical answer given.\n" + "* Another flaw is that the agent me be unfairly penalized for elaborating on the answer, while this benchmark focuses on only and exactly the one canonical answer given.\n", + "* That said, semantic similarity strongly correlates with human judgement of answer quality, so it's a good proxy for evaluating the agent's answers.\n", + " * Source: https://arxiv.org/pdf/2401.17072" ] }, { "cell_type": "code", - "execution_count": 49, + "execution_count": 203, "metadata": {}, "outputs": [], "source": [ - "BENCHMARKS_DIR = \"benchmarks\"\n", - "\n", - "import re\n", - "from string import punctuation\n", - "from nltk.corpus import stopwords\n", - "\n", - "STOP_WORDS = set(stopwords.words('english'))\n", - "\n", "def clean_text(text):\n", " # Lowercase\n", " text = text.lower()\n", @@ -504,9 +481,6 @@ " agent.logger.setLevel(logging.CRITICAL)\n", "\n", " for title, context, question, answer, synthesized_question in tqdm(dfSample.values):\n", - " class Output:\n", - " output: agent_types.AgentType | str = None\n", - "\n", " prompt = synthesized_question\n", " answers_ref.append(answer)\n", " final_answer = agent.run(prompt, stream=False, reset=True)\n", @@ -515,40 +489,50 @@ " answers_ref = [str(answer) for answer in answers_ref]\n", " answers_pred = [str(answer) for answer in answers_pred]\n", "\n", + " dfAnswers = dfSample.copy()\n", + " dfAnswers[\"Predicted Answer\"] = answers_pred\n", + "\n", " # Remove stop words and punctuation from answers\n", " answers_ref = [clean_text(answer) for answer in answers_ref]\n", " answers_pred = [clean_text(answer) for answer in answers_pred]\n", "\n", + " dfAnswers[\"Cleaned Answer\"] = answers_ref\n", + " dfAnswers[\"Cleaned Predicted Answer\"] = answers_pred\n", + "\n", " em = EmbeddingModelWrapper()\n", " similarities = em.get_similarities(\n", " em.get_embeddings( answers_pred ),\n", " em.get_embeddings( answers_ref ),\n", " )\n", "\n", - " dfAnswers = dfSample.copy()\n", - " dfAnswers[\"Predicted Answer\"] = answers_pred\n", " dfAnswers[\"Similarity\"] = similarities\n", "\n", " os.makedirs(BENCHMARKS_DIR, exist_ok=True)\n", - " dfAnswers.to_pickle(os.path.join(BENCHMARKS_DIR, f\"{name}.pkl\"))\n" + " dfAnswers.to_pickle(os.path.join(BENCHMARKS_DIR, f\"{name}.pkl\"))\n", + " return dfAnswers\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "### Set up and run the benchmarks" + "### Set up and run the benchmarks\n", + "\n", + "* We will run the agent with three different prompts:\n", + " * Baseline: The default transformers agent prompt modified only to use the squad_retriever tool.\n", + " * Succinct: The default prompt modified to encourage the agent to be more concise.\n", + " * Focused: The default prompt modified to encourage the agent to focus mostly on SQuAD." ] }, { "cell_type": "code", - "execution_count": 50, + "execution_count": 206, "metadata": {}, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "3681f1eb68ab462fbe06b7000bb34249", + "model_id": "1d96aeac5e244b28a1ec92d3e1ccc115", "version_major": 2, "version_minor": 0 }, @@ -562,7 +546,7 @@ { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "6c10061dff7b454e9b447fdd83a1100a", + "model_id": "98ef07cf016444b59dfc6d9c498d2eed", "version_major": 2, "version_minor": 0 }, @@ -576,7 +560,7 @@ { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "6d9ad710dd1f4474864d6f0d8885d768", + "model_id": "6d2ad4058ee14f858a155338c66c1d91", "version_major": 2, "version_minor": 0 }, @@ -590,7 +574,7 @@ { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "4588847162b344da9fc6524006b42b93", + "model_id": "d38df0a45c2a4514b058f7ce35112757", "version_major": 2, "version_minor": 0 }, @@ -603,61 +587,191 @@ } ], "source": [ + "# Create the agents to be benchmarked\n", "benchmarks = [\n", - " (get_agent(), \"baseline\"), # Baseline agent with default settings\n", - " (get_agent(system_prompt=SUCCINCT_SQUAD_REACT_CODE_SYSTEM_PROMPT), \"succinct\"), # Succinct agent\n", - " (get_agent(system_prompt=FOCUSED_SQUAD_REACT_CODE_SYSTEM_PROMPT), \"focused\"), # Focused agent\n", + " {\"agent\": get_agent(), \"name\": \"baseline\"}, # Baseline agent with default settings\n", + " {\"agent\": get_agent(system_prompt=SUCCINCT_SQUAD_REACT_CODE_SYSTEM_PROMPT), \"name\": \"succinct\"}, # Succinct agent\n", + " {\"agent\": get_agent(system_prompt=FOCUSED_SQUAD_REACT_CODE_SYSTEM_PROMPT), \"name\": \"focused\"}, # Focused agent\n", "]\n", "\n", - "for agent, name in tqdm(benchmarks):\n", - " benchmark_agent(agent, dfSample, name)" + "# Determine if all benchmark files already exist\n", + "benchmark_names = [benchmark[\"name\"] for benchmark in benchmarks]\n", + "all_benchmark_files_exist = all(os.path.exists(os.path.join(BENCHMARKS_DIR, f\"{name}.pkl\")) for name in benchmark_names)\n", + "\n", + "#if benchmark files do not exist, run the benchmarks\n", + "if not all_benchmark_files_exist:\n", + " for benchmark in tqdm(benchmarks):\n", + " benchmark['data'] = benchmark_agent(benchmark[\"agent\"], dfSample, benchmark[\"name\"])\n", + "else:\n", + " # if benchmark files exist, load them\n", + " for benchmark in tqdm(benchmarks):\n", + " benchmark['data'] = pd.read_pickle(os.path.join(BENCHMARKS_DIR, f\"{benchmark['name']}.pkl\"))\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Review the benchmarks for each agent\n", + "\n", + "* We will review the benchmarks for each agent and plot the mean similarity and the distribution of semantic similarity scores across quartiles.\n", + "* We will also plot the number of answers with a semantic similarity score >= a given threshold for each benchmark." ] }, { "cell_type": "code", - "execution_count": 51, + "execution_count": 207, "metadata": {}, "outputs": [ { "data": { - "text/markdown": [ - "## baseline" - ], - "text/plain": [ - "" - ] - }, - "metadata": {}, - "output_type": "display_data" - }, - { - "data": { - "text/markdown": [ - "#### Mean similarity: 0.48" - ], + "image/png": "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", "text/plain": [ - "" + "
" ] }, "metadata": {}, "output_type": "display_data" - }, + } + ], + "source": [ + "# Add mean similarity to each benchmark\n", + "for benchmark in benchmarks:\n", + " benchmark[\"mean_similarity\"] = benchmark[\"data\"][\"Similarity\"].mean()\n", + "\n", + "# Sort benchmarks by mean similarity\n", + "benchmarks.sort(key=lambda x: x[\"mean_similarity\"], reverse=False)\n", + "\n", + "\n", + "# Plot the mean similarity for each benchmark\n", + "mean_similarities = [benchmark[\"mean_similarity\"] for benchmark in benchmarks]\n", + "\n", + "fig, axs = plt.subplots(1, 2, figsize=(12, 6))\n", + "axs[0].bar(benchmark_names, mean_similarities)\n", + "for i, v in enumerate(mean_similarities):\n", + " axs[0].text(i, v + 0.01, str(round(v, 2)), ha=\"center\", va=\"bottom\")\n", + "axs[0].set_xlabel(\"Benchmark\")\n", + "axs[0].set_ylabel(\"Mean Similarity\")\n", + "axs[0].set_title(\"Mean Similarity\")\n", + "\n", + "# Plot the distribution of semantic similarity scores across quartiles\n", + "\n", + "# -- Create a dataframe with the quartile data for all benchmarks combined\n", + "quartiles = [0.25, 0.5, 0.75]\n", + "quartile_data = np.array([])\n", + "quartile_names = np.array([])\n", + "for benchmark in benchmarks:\n", + " df = benchmark[\"data\"]\n", + " semscores = np.array(df[\"Similarity\"].values)\n", + " quartile_data = np.append(quartile_data, np.digitize(semscores, quartiles))\n", + " quartile_names = np.append(quartile_names, [benchmark[\"name\"]] * len(semscores))\n", + "\n", + "df = pd.DataFrame({\"name\": quartile_names, \"quartile\": quartile_data})\n", + "\n", + "# -- Plot the distribution of semantic similarity scores across quartiles\n", + "hue_order = list(df[\"quartile\"].unique()) # Best performers on top\n", + "hue_order.sort(reverse=True)\n", + "ax = sns.histplot(\n", + " df,\n", + " x=\"name\",\n", + " hue=\"quartile\",\n", + " multiple=\"stack\",\n", + " hue_order=hue_order,\n", + " palette=[\"#a2d9a4\", \"#47a0b3\", \"#fca55d\", \"#e2514a\"],\n", + ")\n", + "ax.set_xlabel(\"Benchmark\")\n", + "ax.set_ylabel(\"Quartile\")\n", + "ax.set_title(\"Distribution of Semantic Similarity Scores\")\n", + "ax.legend(\n", + " title=\"Quartile\", labels=[\"Poor\", \"Needs Improvement\", \"Acceptable\", \"Excellent\"]\n", + ")\n", + "\n", + "# -- Add the counts to the bars for easy reference\n", + "for container in ax.containers:\n", + " labels = [\n", + " f\"{round(v.get_height())}\" if v.get_height() > 0 else \"\" for v in container\n", + " ]\n", + " ax.bar_label(container, labels=labels, label_type=\"center\")\n", + "\n", + "plt.tight_layout()\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Observations\n", + "\n", + "* The `succinct` agent represents a significant improvement over the baseline, with a mean similarity of `0.83` and a distribution of semantic similarity scores across quartiles that is much closer to the top performer.\n", + "* However, the `focused` agent out-performs the `succinct` agent, with a mean similarity of `0.85`, with a handful of better answers.\n", + "* The performance is close enough that it may be worth while to look at the number of answers that would be considered correct across a range of possible thresholds. \n", + "\n", + "#### Number of answers with a semantic similarity score >= a given threshold\n", + "\n", + "* Note: that I show every possible threshold starting at 0.01, but in practice it's unlikely that a threshold of 0.01 would be used as a threshold for acceptable answers." + ] + }, + { + "cell_type": "code", + "execution_count": 208, + "metadata": {}, + "outputs": [ { "data": { - "text/markdown": [ - "#### Number of rows with similarity score less than 0.9: 98" - ], + "image/png": "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", "text/plain": [ - "" + "
" ] }, "metadata": {}, "output_type": "display_data" - }, + } + ], + "source": [ + "# Plot the number of answers with a semantic similarity score >= a given threshold for each benchmark\n", + "for benchmark in benchmarks:\n", + " thresholds = np.arange(0.01, 0.99, 0.01)\n", + " num_rows_above_threshold = []\n", + " for threshold in thresholds:\n", + " num_rows_above_threshold.append(len(benchmark['data'][benchmark['data']['Similarity'] >= threshold]))\n", + " benchmark['num_rows_above_threshold'] = num_rows_above_threshold\n", + "\n", + "for benchmark in benchmarks:\n", + " plt.plot(thresholds, benchmark['num_rows_above_threshold'], label=benchmark['name'])\n", + "plt.xlabel('Threshold')\n", + "plt.ylabel('Number of answers')\n", + "plt.title('Number of answers with semantic similarity score >= threshold')\n", + "plt.legend()\n", + "plt.show()\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Observations\n", + "\n", + "* The `focused` agent outperforms the `succinct` agent by a small margin across most thresholds, until about 0.7, where it clearly outperforms the `succinct` agent.\n", + "* Given this, I will deliver the `focused` agent as the best performing agent." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Exactly correct answers" + ] + }, + { + "cell_type": "code", + "execution_count": 218, + "metadata": {}, + "outputs": [ { "data": { "text/markdown": [ - "#### Root Mean Squared Error: 0.57" + "#### Number of answers with Similarity == 1.0 for baseline" ], "text/plain": [ "" @@ -668,11 +782,8 @@ }, { "data": { - "text/markdown": [ - "#### Root Mean Squared Log Error: 0.44" - ], "text/plain": [ - "" + "0" ] }, "metadata": {}, @@ -681,7 +792,7 @@ { "data": { "text/markdown": [ - "## succinct" + "#### Number of answers with Similarity == 1.0 for succinct" ], "text/plain": [ "" @@ -692,11 +803,8 @@ }, { "data": { - "text/markdown": [ - "#### Mean similarity: 0.83" - ], "text/plain": [ - "" + "42" ] }, "metadata": {}, @@ -705,7 +813,7 @@ { "data": { "text/markdown": [ - "#### Number of rows with similarity score less than 0.9: 37" + "#### Number of answers with Similarity == 1.0 for focused" ], "text/plain": [ "" @@ -716,20 +824,39 @@ }, { "data": { - "text/markdown": [ - "#### Root Mean Squared Error: 0.31" - ], "text/plain": [ - "" + "45" ] }, "metadata": {}, "output_type": "display_data" - }, + } + ], + "source": [ + "# Show how many answers have Similarity == 1.0 exactly\n", + "for benchmark in benchmarks:\n", + " display(Markdown(f\"#### Number of answers with Similarity == 1.0 for {benchmark['name']}\"))\n", + " display(benchmark['data'][benchmark['data']['Similarity'] == 1.0].shape[0])" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Observations\n", + "\n", + "* It's impressive that the `focused` and `succinct` agents both got nearly 50% of the answers exactly correct." + ] + }, + { + "cell_type": "code", + "execution_count": 214, + "metadata": {}, + "outputs": [ { "data": { "text/markdown": [ - "#### Root Mean Squared Log Error: 0.24" + "#### Worse scoring answers for baseline" ], "text/plain": [ "" @@ -740,11 +867,104 @@ }, { "data": { - "text/markdown": [ - "## focused" + "text/html": [ + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
 Synthesized QuestionAnswerPredicted AnswerSimilarity
31138How many species of fungi, both non-lichen-forming and lichen-forming, have been recorded in Antarctica?1150About 1150 species of fungi have been recorded in Antarctica, including approximately 750 non-lichen-forming and 400 lichen-forming species.-0.037171
56941What fraction of the South Florida population lives in the city of Miami, given that it is home to less than one-thirteenth of the population of South Florida?one-thirteenthApproximately 7.69% of the South Florida population lives in the city of Miami.0.011993
31553In how many scenarios projected by the ABS will Sydney remain higher than Melbourne in population beyond 2056?twoIn two scenarios projected by the ABS, Sydney will remain higher than Melbourne in population beyond 2056.0.075540
59122How many companies were involved in the development of USB in 1994?sevenSeven companies were involved in the development of USB in 1994: Compaq, DEC, IBM, Intel, Microsoft, NEC, and Nortel.0.107220
41377How many Freistaaten (Free States) are there in Germany according to the current political structure?threeThere are 16 Freistaaten in Germany according to the current political structure.0.131161
20149How large is the Marshall Islands shark sanctuary in square miles, specifically referring to the nearly 2,000,000 square kilometers designated by the government?772,000The Marshall Islands shark sanctuary is approximately 772,204 square miles in size.0.135708
51916What kind of nutritional value do processed foods have in comparison to their fresh variants, considering the impact of processing techniques on nutrient content as described in the context of nutrition?reducedProcessed foods generally have a reduced nutritional value compared to fresh foods, which are nutritionally superior. Processing techniques can lead to the loss of essential nutrients and may introduce harmful substances.0.140512
83184In March 2014, what was the number of people sentenced to death during the single hearing by the Minya Criminal Court in Egypt?529I could not find the number of people sentenced to death during the single hearing by the Minya Criminal Court in Egypt in March 2014.0.140897
16329On what date did ESPN announce the purchase of a minority stake in the Arena Football League (AFL)?December 19, 2006ESPN announced the purchase of a minority stake in the Arena Football League (AFL) on December 19, 2006.0.144917
6468What does the Sanskrit term \"Karma,\" which translates to \"action\" or \"work,\" refer to in the context of Buddhism?action, workIn Buddhism, the Sanskrit term 'Karma' refers to the force that drives saṃsāra, the cycle of suffering and rebirth. It encompasses actions of body, speech, or mind that arise from mental intent and produce consequences, with good actions leading to positive outcomes and bad actions leading to negative consequences.0.168453
\n" ], "text/plain": [ - "" + "" ] }, "metadata": {}, @@ -753,7 +973,7 @@ { "data": { "text/markdown": [ - "#### Mean similarity: 0.86" + "#### Worse scoring answers for succinct" ], "text/plain": [ "" @@ -764,11 +984,104 @@ }, { "data": { - "text/markdown": [ - "#### Number of rows with similarity score less than 0.9: 31" + "text/html": [ + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
 Synthesized QuestionAnswerPredicted AnswerSimilarity
31138How many species of fungi, both non-lichen-forming and lichen-forming, have been recorded in Antarctica?11501150 species of fungi have been recorded in Antarctica.-0.002039
79931Where was very expensive wallpaper imported from during the Georgian architectural period?ChinaThe information about the origin of very expensive wallpaper during the Georgian architectural period is not available in the dataset.0.165669
60191According to Hayek, limited government power through the Rule of Law does not stultify individual efforts by ad hoc action. What does this mean for people's ability to make investments and future plans?frustrate his effortsAccording to Hayek, limited government power through the Rule of Law enables individuals to confidently make investments and future plans without fear of arbitrary government interference.0.172209
51916What kind of nutritional value do processed foods have in comparison to their fresh variants, considering the impact of processing techniques on nutrient content as described in the context of nutrition?reducedProcessed foods have a reduced nutritional value compared to fresh foods.0.174138
6468What does the Sanskrit term \"Karma,\" which translates to \"action\" or \"work,\" refer to in the context of Buddhism?action, workIn Buddhism, Karma refers to the force that drives saṃsāra, encompassing actions of body, speech, or mind that stem from mental intent and produce consequences.0.266618
49374Who is elected as mayor every even-numbered year in Ann Arbor, Michigan?mayorChristopher Taylor0.288215
69135How many engineering colleges are there in Rajasthan, as mentioned in the context provided?41 engineering colleges410.322794
5111When did Tajiks begin to be conscripted into the Soviet Army, particularly during the lead-up to and including World War II?1939Tajiks began to be conscripted into the Soviet Army in 1939.0.329918
13080What impact does temperature have on the variability of hunter-gatherer tool kits?increased variability of toolsTemperature increases the variability of hunter-gatherer tool kits.0.341771
69990When was ZE Records, the no wave-affiliated label mentioned in the context of post-punk, founded?1978ZE Records was founded in 1978. It is known for its association with the no wave movement and its influence on the post-punk music scene.0.343213
\n" ], "text/plain": [ - "" + "" ] }, "metadata": {}, @@ -777,7 +1090,7 @@ { "data": { "text/markdown": [ - "#### Root Mean Squared Error: 0.27" + "#### Worse scoring answers for focused" ], "text/plain": [ "" @@ -788,11 +1101,104 @@ }, { "data": { - "text/markdown": [ - "#### Root Mean Squared Log Error: 0.21" + "text/html": [ + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
 Synthesized QuestionAnswerPredicted AnswerSimilarity
31138How many species of fungi, both non-lichen-forming and lichen-forming, have been recorded in Antarctica?11501150 species of fungi have been recorded in Antarctica, including 750 non-lichen-forming and 400 lichen-forming species.-0.034279
51916What kind of nutritional value do processed foods have in comparison to their fresh variants, considering the impact of processing techniques on nutrient content as described in the context of nutrition?reducedProcessed foods have reduced nutritional value compared to fresh foods.0.174138
56941What fraction of the South Florida population lives in the city of Miami, given that it is home to less than one-thirteenth of the population of South Florida?one-thirteenthless than 1/130.192653
6468What does the Sanskrit term \"Karma,\" which translates to \"action\" or \"work,\" refer to in the context of Buddhism?action, workIn Buddhism, 'Karma' refers to the actions of body, speech, or mind that spring from mental intent and drive the cycle of suffering and rebirth (saṃsāra).0.245145
60191According to Hayek, limited government power through the Rule of Law does not stultify individual efforts by ad hoc action. What does this mean for people's ability to make investments and future plans?frustrate his effortsHayek believes that limited government power through the Rule of Law allows individuals to make wise investments and future plans with confidence, as it prevents the government from frustrating their efforts.0.251189
79931Where was very expensive wallpaper imported from during the Georgian architectural period?ChinaVery expensive wallpaper during the Georgian architectural period was primarily imported from France and China.0.277804
49374Who is elected as mayor every even-numbered year in Ann Arbor, Michigan?mayorChristopher Taylor0.288215
5111When did Tajiks begin to be conscripted into the Soviet Army, particularly during the lead-up to and including World War II?1939Tajiks began to be conscripted into the Soviet Army in 1939.0.329918
13080What impact does temperature have on the variability of hunter-gatherer tool kits?increased variability of toolsTemperature increases the variability of hunter-gatherer tool kits.0.341771
45263How many copies has Queen's Greatest Hits II sold worldwide?16 millionThe specific sales figures for Queen's Greatest Hits II are not available, but Queen has sold over 150 million records worldwide in total.0.345838
\n" ], "text/plain": [ - "" + "" ] }, "metadata": {}, @@ -800,92 +1206,40 @@ } ], "source": [ - "# Load and display all benchmarks\n", - "def load_benchmarks():\n", - " benchmarks_dir = \"benchmarks\"\n", - " benchmarks = []\n", - " for file in os.listdir(benchmarks_dir):\n", - " if file.endswith(\".pkl\"):\n", - " df = pd.read_pickle(os.path.join(benchmarks_dir, file))\n", - " benchmarks.append({\n", - " 'name': file.replace(\".pkl\", \"\"),\n", - " 'data': df, \n", - " })\n", - " return benchmarks\n", - "\n", - "benchmarks = load_benchmarks()\n", - "\n", - "# Add mean similarity to each benchmark\n", - "for benchmark in benchmarks:\n", - " benchmark['mean_similarity'] = benchmark['data']['Similarity'].mean()\n", - "\n", - "# Sort benchmarks by mean similarity\n", - "benchmarks.sort(key=lambda x: x['mean_similarity'], reverse=False)\n", - "\n", - "from sklearn.metrics import roc_curve\n", + "# Show the worst 10 answers for each benchmark\n", "for benchmark in benchmarks:\n", - " display(Markdown(f\"## {benchmark['name']}\"))\n", - " display(Markdown(f\"#### Mean similarity: {round(benchmark['mean_similarity'], 2)}\"))\n", - "\n", - " # Count the number of rows where the similarity score is less than 0.9\n", - " num_low_similarity = len(benchmark['data'][benchmark['data']['Similarity'] < 0.9])\n", - " display(Markdown(f\"#### Number of rows with similarity score less than 0.9: {num_low_similarity}\"))\n", - " # df = benchmark['data'][benchmark['data']['Similarity'] < 0.9]\n", - " # display_text_df(df[['Synthesized Question', 'Answer', 'Predicted Answer', 'Similarity']])\n", - "\n", - " # For thresholds from 0.5 to 0.99, count the number of rows where the similarity score is less than the threshold\n", - " thresholds = np.arange(0.1, 0.9, 0.01)\n", - " num_rows_below_threshold = []\n", - " for threshold in thresholds:\n", - " num_rows_below_threshold.append(len(benchmark['data'][benchmark['data']['Similarity'] < threshold]))\n", - " benchmark['num_rows_below_threshold'] = num_rows_below_threshold\n", - "\n", - " # Considering 1-Similarity as error\n", - " error = 1 - benchmark['data']['Similarity']\n", - " # Calculate Root Mean Squared Error(RMSE)\n", - " rmse = np.sqrt(np.mean(np.square(error)))\n", - " display(Markdown(f\"#### Root Mean Squared Error: {round(rmse, 2)}\"))\n", - " # Calculate Root Mean Squared Log Error(RMSLE)\n", - " rmsle = np.sqrt(np.mean(np.square(np.log1p(error))))\n", - " display(Markdown(f\"#### Root Mean Squared Log Error: {round(rmsle, 2)}\"))" + " display(Markdown(f\"#### Worse scoring answers for {benchmark['name']}\"))\n", + " display_text_df(\n", + " benchmark[\"data\"][\n", + " [\"Synthesized Question\", \"Answer\", \"Predicted Answer\", \"Similarity\"]\n", + " ]\n", + " .sort_values(by=\"Similarity\", ascending=True)\n", + " .head(10)\n", + " )" ] }, { - "cell_type": "code", - "execution_count": 53, + "cell_type": "markdown", "metadata": {}, - "outputs": [ - { - "data": { - "image/png": "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", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], "source": [ - "# Plot the number of rows below each threshold, for each benchmark on the same plot\n", - "import matplotlib.pyplot as plt\n", - "for benchmark in benchmarks:\n", - " plt.plot(thresholds, benchmark['num_rows_below_threshold'], label=benchmark['name'])\n", - "plt.xlabel('Threshold')\n", - "plt.ylabel('Number of answers with semantic similarity score less than threshold')\n", - "plt.legend()\n", - "plt.show()\n" + "#### Observations\n", + "\n", + "* All of the agents are unfairly penalized for elaborating on answers when the reference answer is brief. \n", + " * For example, in the worst-ranking answer when all 3 agents are correct (id=31138), but ranked the worst in semantic similarity due to the extra context they provided.\n", + "* Of the `baseline` agent's worst answers by semantic similarity, 3 are unquestionably incorrect, while the rest look correct from a human evaluator's perspective.\n", + "* Of the `succinct` agent's worst answers by semantic similarity, 2 are unquestionably incorrect, while the rest look correct from a human evaluator's perspective.\n", + "* Of the `focused` agent's worst answers by semantic similarity, 2 are unquestionably incorrect, while the rest look correct from a human evaluator's perspective." ] }, { "cell_type": "code", - "execution_count": 56, + "execution_count": 215, "metadata": {}, "outputs": [ { "data": { "text/markdown": [ - "## Worse scoring answers for baseline" + "#### Best scoring answers for baseline" ], "text/plain": [ "" @@ -898,91 +1252,102 @@ "data": { "text/html": [ "\n", - "\n", + "
\n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", "
Synthesized QuestionCleaned AnswerCleaned Predicted AnswerSimilarity Synthesized QuestionAnswerPredicted AnswerSimilarity
How many species of fungi, both non-lichen-forming and lichen-forming, have been recorded in Antarctica?11501150 species fungi recorded antarctica including approximately 750 nonlichenforming 400 lichenforming species-0.03717156283While Darwin wasn't sure about blending inherited traits, what was his opinion on the usefulness of inherited variation in relation to his concept of selection as discussed in \"On the Origin of Species\"?Darwin's concept of selection working on a population with a range of small variations was workableDarwin believed that inherited variation was useful because his concept of selection could work on a population with a range of small variations.0.830038
What fraction of the South Florida population lives in the city of Miami, given that it is home to less than one-thirteenth of the population of South Florida?one-thirteenthfraction south florida population lives city miami less 113-0.00369715471What is the name of the main public service broadcaster in Somalia, as mentioned in the context of Communications in Somalia?Somali National TelevisionThe main public service broadcaster in Somalia is Somali National Television (SNTV).0.816376
In how many scenarios projected by the ABS will Sydney remain higher than Melbourne in population beyond 2056?twotwo scenarios projected abs sydney remain higher melbourne population beyond 20560.07554039287What specific educational reforms implemented by Napoleon contributed to the establishment of a modernized education system in France and influenced education across Europe?Napoleon's educational reformsNapoleon implemented several educational reforms that contributed to a modernized education system in France and influenced Europe. These reforms included the establishment of state secondary schools (lycées) to provide standardized education, the emphasis on teaching sciences and languages, and the development of a secular public education system. He aimed to synthesize the best elements from the Ancien Régime, the Enlightenment, and the Revolution, and his system featured scholarships and strict discipline, resulting in France's educational system outperforming its European counterparts.0.810252
How many companies were involved in the development of USB in 1994?sevenseven companies involved development usb 19940.10881436802Who is responsible for approving drugs for use in the United Kingdom, according to the context of the pharmaceutical industry?Medicines and Healthcare Products Regulatory AgencyThe Medicines and Healthcare Products Regulatory Agency (MHRA) is responsible for approving drugs for use in the United Kingdom.0.810035
What kind of nutritional value do processed foods have in comparison to their fresh variants, considering the impact of processing techniques on nutrient content as described in the context of nutrition?reducedprocessed foods generally reduced nutritional value compared fresh foods superior nutrient content processing techniques lead loss essential nutrients often result addition harmful substances0.13896556292How does Darwin theorize that instincts have evolved in slave-making ants in Chapter VII of \"On the Origin of Species\"?He thought it reasonable that species with an extreme dependency on slave workers had evolved in incremental steps.Darwin theorized that instincts in slave-making ants evolved in incremental steps, particularly in species that exhibit extreme dependency on slave workers.0.809778
On what date did ESPN announce the purchase of a minority stake in the Arena Football League (AFL)?December 19, 2006espn announced purchase minority stake arena football league december 19 20060.14332174315What is the name of the famous valley located in Oregon, specifically mentioned in the context of the Geography of the United States?Willamette ValleyThe famous valley located in Oregon is the Willamette Valley.0.800726
What term did Europeans use to refer to the Ottoman Empire during the mid-19th century?the \"sick man\"term europeans used refer ottoman empire mid19th century sick man europe0.14857532186What did the Observer Corps in Britain observe and report on regarding hostile aircraft flying over Britain?hostile aircraft flying over BritainThe Observer Corps in Britain observed and reported on hostile aircraft flying over Britain.0.787479
What does the Sanskrit term \"Karma,\" which translates to \"action\" or \"work,\" refer to in the context of Buddhism?action, workbuddhism sanskrit term karma refers force drives saṃsāra cycle suffering rebirth encompasses actions body speech mind arise mental intent lead consequences either life future rebirths good actions encouraged unwholesome actions avoided0.16040969135How many engineering colleges are there in Rajasthan, as mentioned in the context provided?41 engineering collegesThere are 41 engineering colleges in Rajasthan.0.774651
How many civilians died in the attack on the Vilnius TV Tower during the events leading up to the dissolution of the Soviet Union?Fourteenfourteen unarmed civilians died attack vilnius tv tower events leading dissolution soviet union0.16408670472What factors contributed to Prussia's decision not to renew its alliance with Britain after the Seven Years' War?Prussians were angered by what they considered a British betrayal in 1762Prussia decided not to renew its alliance with Britain after the Seven Years' War primarily due to their anger over what they considered a British betrayal in 1762.0.772004
At around what age, specifically twenty-four or twenty-five, was Dominic ordained as a Priest in the context of the Dominican Order?twenty-fivedominic ordained priest age twentyfive0.16567281021Where were the advanced materials and techniques, such as buttresses and arches, used in Sumerian temples and palaces on display?Sumerian temples and palacesThe advanced materials and techniques, such as buttresses and arches, used in Sumerian temples and palaces are on display in Sumerian temples and palaces.0.766647
\n" ], "text/plain": [ - "" + "" ] }, "metadata": {}, @@ -991,7 +1356,7 @@ { "data": { "text/markdown": [ - "## Worse scoring answers for succinct" + "#### Best scoring answers for succinct" ], "text/plain": [ "" @@ -1004,91 +1369,102 @@ "data": { "text/html": [ "\n", - "\n", + "
\n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", "
Synthesized QuestionCleaned AnswerCleaned Predicted AnswerSimilarity Synthesized QuestionAnswerPredicted AnswerSimilarity
How many species of fungi, both non-lichen-forming and lichen-forming, have been recorded in Antarctica?11501150 species fungi recorded antarctica-0.00203955901In 1849, what railroad was created by the North Carolina legislature to extend west from the Wilmington and Raleigh Railroad?North Carolina RailroadNorth Carolina Railroad1.000000
Where was very expensive wallpaper imported from during the Georgian architectural period?Chinacould find information origins expensive wallpaper georgian architectural period0.15451534182What ordering scheme did readers prefer during the Age of Enlightenment?alphabeticalalphabetical1.000000
What kind of nutritional value do processed foods have in comparison to their fresh variants, considering the impact of processing techniques on nutrient content as described in the context of nutrition?reducedprocessed foods reduced nutritional value compared fresh foods0.1741389983What year was the Banská Akadémia, the world's first institution of technology, founded in Banská Štiavnica, Slovakia?173517351.000000
According to Hayek, limited government power through the Rule of Law does not stultify individual efforts by ad hoc action. What does this mean for people's ability to make investments and future plans?frustrate his effortsaccording hayek limited government power rule law enables individuals confidently make investments future plans without fear government interference could frustrate efforts0.21299617172What year did Bern join the Swiss Confederacy, according to the historical context provided?135313531.000000
What does the Sanskrit term \"Karma,\" which translates to \"action\" or \"work,\" refer to in the context of Buddhism?action, workbuddhism karma refers force drives saṃsāra encompassing actions body speech mind spring mental intent producing consequences0.27754287251Where can safari hunters go that are considered uninviting to typical ecotourists, particularly in the context of hunting in Tanzania?remote areasremote areas1.000000
Who is elected as mayor every even-numbered year in Ann Arbor, Michigan?mayorchristopher taylor0.28821565484What field did the majority of the 144,600 employees at La Défense work in as of 2010?finance and insurancefinance and insurance1.000000
How many copies has Queen's Greatest Hits II sold worldwide?16 millionsales figures queens greatest hits ii specifically available queen sold 150 million records worldwide0.31570474315What is the name of the famous valley located in Oregon, specifically mentioned in the context of the Geography of the United States?Willamette ValleyWillamette Valley1.000000
When did Tajiks begin to be conscripted into the Soviet Army, particularly during the lead-up to and including World War II?1939tajiks began conscripted soviet army 19390.32991843267What is another speed rating that can also be reported by the camera in addition to the noise-based and saturation-based speeds?SOS-based speedSOS-based speed1.000000
What impact does temperature have on the variability of hunter-gatherer tool kits?increased variability of toolstemperature increases variability huntergatherer tool kits0.34177116329On what date did ESPN announce the purchase of a minority stake in the Arena Football League (AFL)?December 19, 2006December 19, 20061.000000
What was the typical shape and size of town terraced houses during the Georgian architecture period?tall and narrowtown terraced houses georgian architecture period typically tall narrow0.35948058193At around what age, specifically twenty-four or twenty-five, was Dominic ordained as a Priest in the context of the Dominican Order?twenty-fivetwenty-five1.000000
\n" ], "text/plain": [ - "" + "" ] }, "metadata": {}, @@ -1097,7 +1473,7 @@ { "data": { "text/markdown": [ - "## Worse scoring answers for focused" + "#### Best scoring answers for focused" ], "text/plain": [ "" @@ -1110,91 +1486,102 @@ "data": { "text/html": [ "\n", - "\n", + "
\n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " \n", + " \n", " \n", " \n", "
Synthesized QuestionCleaned AnswerCleaned Predicted AnswerSimilarity Synthesized QuestionAnswerPredicted AnswerSimilarity
What kind of nutritional value do processed foods have in comparison to their fresh variants, considering the impact of processing techniques on nutrient content as described in the context of nutrition?reducedprocessed foods generally reduced nutritional value compared fresh foods0.15657834182What ordering scheme did readers prefer during the Age of Enlightenment?alphabeticalalphabetical1.000000
Where was very expensive wallpaper imported from during the Georgian architectural period?Chinageorgian architectural period expensive wallpaper primarily imported france china french wallpapers particularly made intricate designs luxurious materials highly sought wealthy elite britain additionally chinese wallpapers often featuring elaborate patterns rich colors became popular time well importation wallpapers reflected trends interior design desire opulence among upper classes georgian society0.2066123951In 1860, approximately how many Irish immigrants were living in New York City?200,000Over 200,0001.000000
According to Hayek, limited government power through the Rule of Law does not stultify individual efforts by ad hoc action. What does this mean for people's ability to make investments and future plans?frustrate his effortsaccording hayek limited government power rule law allows individuals make wise investments future plans confidence prevents government frustrating efforts0.21823855901In 1849, what railroad was created by the North Carolina legislature to extend west from the Wilmington and Raleigh Railroad?North Carolina RailroadNorth Carolina Railroad1.000000
What does the Sanskrit term \"Karma,\" which translates to \"action\" or \"work,\" refer to in the context of Buddhism?action, workbuddhism karma refers force drives saṃsāra involving actions body speech mind spring mental intent produce consequences0.28260581694When did the United States purchase Alaska from Russia?186718671.000000
Who is elected as mayor every even-numbered year in Ann Arbor, Michigan?mayorchristopher taylor0.2882156266Where was Donda West's funeral held on November 20, 2007?Oklahoma CityOklahoma City1.000000
What impact does temperature have on the variability of hunter-gatherer tool kits?increased variability of toolstemperature leads increased variability huntergatherer tool kits0.3140882267How many households were the offices of Qianhu in charge of during the Ming dynasty as described in the context of Sino-Tibetan relations?1,000 households1,000 households1.000000
When did Tajiks begin to be conscripted into the Soviet Army, particularly during the lead-up to and including World War II?1939tajiks began conscripted soviet army 19390.32991972912What year did the government start distributing Morrison shelters during The Blitz?194119411.000000
What are the implications of measuring heterosexuality and homosexuality on separate scales instead of the same scale, as discussed in the context of the Kinsey scale in the study of sexual orientation?they act as tradeoffs such, whereby to be more feminine one had to be less masculine and vice versameasuring heterosexuality homosexuality separate scales allows independent determination orientations enabling individuals identify heterosexual homosexual simultaneously0.33065640855Who was the chief engineer at the United States Electric Lighting Company, as mentioned in the context of the development of the incandescent light bulb?Hiram S. MaximHiram S. Maxim1.000000
What is the size of the Matthaei Botanical Gardens located in Ann Arbor, Michigan?300 acresmatthaei botanical gardens ann arbor michigan covers area approximately 300 acres0.3856608786What was Alfred North Whitehead's final area of study before developing his comprehensive metaphysical system?metaphysicsmetaphysics1.000000
What term did Europeans use to refer to the Ottoman Empire during the mid-19th century?the \"sick man\"sick man europe0.45129458193At around what age, specifically twenty-four or twenty-five, was Dominic ordained as a Priest in the context of the Dominican Order?twenty-fivetwenty-five1.000000
\n" ], "text/plain": [ - "" + "" ] }, "metadata": {}, @@ -1202,15 +1589,33 @@ } ], "source": [ - "# Show the worst 10 answers for each benchmark\n", + "# Show the best 10 answers for each benchmark\n", "for benchmark in benchmarks:\n", - " display(Markdown(f\"## Worse scoring answers for {benchmark['name']}\"))\n", - " # Rename columns\n", - " df = benchmark['data'].rename(columns={\n", - " 'Answer': 'Cleaned Answer',\n", - " 'Predicted Answer': 'Cleaned Predicted Answer'\n", - " })\n", - " display_text_df(df[['Synthesized Question', 'Cleaned Answer', 'Cleaned Predicted Answer', 'Similarity']].sort_values(by='Similarity', ascending=True).head(10))\n" + " display(Markdown(f\"#### Best scoring answers for {benchmark['name']}\"))\n", + " display_text_df(benchmark['data'][['Synthesized Question', 'Answer', 'Predicted Answer', 'Similarity']].sort_values(by='Similarity', ascending=False).head(10))" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Observations\n", + "\n", + "* Here we can see `focused` and `succeinct` agents are producing exactly correct answers, while even at its best, the `baseline` agent is producing factually relevant and correct answers, but being penalized for elaborating and providing more context. " + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## BenchmarkingConclusion\n", + "\n", + "Overall, the semantic similarity metric differentiates between good and bad answers, but is prone to penalizing agents for elaborating on answers when the reference answer is brief. \n", + "\n", + "## BenchmarkingFuture Work\n", + "\n", + "* It may be interesting to establish both concise and contextualized acceptable answers, and then take the max sementic similarity score between the predicted answer and the acceptable answers.\n", + "* It would also be interesting to look for and include the other acceptable answers found in the SQuAD dataset. " ] } ],