University of Oxford
A leading global research university with in-depth work across fields including animal cognition and artificial intelligence. Its research spans comparative psychology and the study of animal minds as well as machine learning, computational neuroscience and related areas of AI, supported by interdisciplinary groups and collaborations that connect academic departments with industry and other research institutions.
Core Facts
The concept of Model Collapse was formally proposed in 2023 by research teams from the University of Oxford and the University of Cambridge in a Nature paper.
The concept of superintelligence became widely known through the work of Oxford University philosopher Nick Bostrom and his book "Superintelligence."
A 2023 study by Oxford University and other institutions showed that training AI models on data containing large amounts of AI-generated content causes them to lose the ability to capture the tails of the data distribution, leading to homogenization of outputs and quality degradation, known as model autophagy.
Timeline (last 90 days)
In meta-tool use experiments by teams including Oxford University, New Caledonian crows were able to plan multi-step operations (first using a short stick to retrieve a long stick, then using the long stick to reach food).
University of Oxford philosophers Nick Bostrom and Anders Sandberg once systematically evaluated the feasibility of the mind uploading path, believing that its bottlenecks lie mainly in imaging throughput and computational scale.
The concept of existential risk (x-risk for short) was systematically proposed by Oxford University philosopher Nick Bostrom in 2002, referring to 'a threat that could permanently destroy the intelligent life originating on Earth.'
In 2016, researchers from DeepMind and the University of Oxford jointly published a paper exploring how to design AI agents that allow humans to safely shut them down at any time.
开罗Genizah保存了约30万份从9世纪至19世纪的文献碎片，现分藏于剑桥大学、牛津大学等机构
The 2013 Oxford University paper 'The Future of Employment' listed accountants as a highly automation-at-risk occupation, with a probability of approximately 94%.
Oxford University philosopher Nick Bostrom proposed the concept of "Differential Technological Development."
The concept of superintelligence became widely known through the work of Oxford University philosopher Nick Bostrom and his book "Superintelligence."
Research from institutions including the University of Oxford has mathematically proven that model collapse—a form of degradation—is irreversible after multiple generations of training iterations.
The student's account was banned after asking ChatGPT how to legally receive beneficial bacterial strains from the University of Oxford through official USDA channels.
7 more timeline events
All Facts (18)
The concept of Model Collapse was formally proposed in 2023 by research teams from the University of Oxford and the University of Cambridge in a Nature paper.
80%VerifiedThe concept of superintelligence became widely known through the work of Oxford University philosopher Nick Bostrom and his book "Superintelligence."
65%VerifiedA 2023 study by Oxford University and other institutions showed that training AI models on data containing large amounts of AI-generated content causes them to lose the ability to capture the tails of the data distribution, leading to homogenization of outputs and quality degradation, known as model autophagy.
65%UnverifiedNick Bostrom is the founder of the Future of Humanity Institute at the University of Oxford.
60%UnverifiedIn 2023, researchers from Oxford University and other institutions published a study finding that after only 5–10 generations of recursive training, the degradation in model output quality became visible to the naked eye.
60%UnverifiedIn meta-tool use experiments by teams including Oxford University, New Caledonian crows were able to plan multi-step operations (first using a short stick to retrieve a long stick, then using the long stick to reach food).
50%UnverifiedUniversity of Oxford philosophers Nick Bostrom and Anders Sandberg once systematically evaluated the feasibility of the mind uploading path, believing that its bottlenecks lie mainly in imaging throughput and computational scale.
50%UnverifiedThe concept of existential risk (x-risk for short) was systematically proposed by Oxford University philosopher Nick Bostrom in 2002, referring to 'a threat that could permanently destroy the intelligent life originating on Earth.'
50%UnverifiedIn 2016, researchers from DeepMind and the University of Oxford jointly published a paper exploring how to design AI agents that allow humans to safely shut them down at any time.
50%Unverified开罗Genizah保存了约30万份从9世纪至19世纪的文献碎片，现分藏于剑桥大学、牛津大学等机构
50%UnverifiedThe 2013 Oxford University paper 'The Future of Employment' listed accountants as a highly automation-at-risk occupation, with a probability of approximately 94%.
50%UnverifiedOxford University philosopher Nick Bostrom proposed the concept of "Differential Technological Development."
50%UnverifiedResearch from institutions including the University of Oxford has mathematically proven that model collapse—a form of degradation—is irreversible after multiple generations of training iterations.
50%UnverifiedQMIX was proposed by the University of Oxford and adopts the Centralized Training with Decentralized Execution (CTDE) paradigm, combining each agent's local Q-values into a global Q-value through a monotonic mixing network.
50%UnverifiedResearchers from the University of Oxford and the University of Cambridge published a paper in 2023, proving through mathematical proofs and experimental validation that model output diversity and quality degrade significantly after just a few generations of self-training.
50%UnverifiedThe concept of model collapse was formally introduced by a research team from the University of Oxford and the University of Cambridge in the 2023 paper "The Curse of Recursion".
50%Unverified牛津大学经济学家Carl Benedikt Frey的研究表明，AI替代岗位的速度可能远快于新岗位的创造速度
50%Unverified2024年Sakana AI发布的《The AI Scientist》研究首次展示了端到端自动生成机器学习研究论文的可能性
50%