Yep Moravec's Paradox is highly related
RESEARCH
-
LeanDojo: Open-Source Framework for Lean and LLMs
By
–
https://t.co/yYKUrhx7UG was the first open-source framework that combined Lean with LLMs. We have also released Lean co-pilot to help you write proofs in Lean. The benefit of Lean is 100% accuracy and no hallucination in generated proofs. https://t.co/CWwYulnvFc https://t.co/BMxK3cio1l
— Prof. Anima Anandkumar (@AnimaAnandkumar) 25 juillet 2024https://
leandojo.org was the first open-source framework that combined Lean with LLMs. We have also released Lean co-pilot to help you write proofs in Lean. The benefit of Lean is 100% accuracy and no hallucination in generated proofs. https://
github.com/lean-dojo/Lean
Copilot
… -
Neuro-symbolic AI: Language Models and Symbolic Reasoning
By
–
Neuro-symbolic paradigm strikes again! If in #AlphaGeometry, the creative language model (System 1) suggests insights for the reliable symbolic engine (System 2) to complete a proof, we see that pattern again in #AlphaProof. The language model suggests key proof steps in a
-
LLMs Imitate Training Data Format During Post-Training Stage
By
–
It does that because all of its training data in the last, post-training stage are of the form [question -> authoritative sounding solution], where the solutions are written by humans. The LLMs just imitate the form/style of that training data.
-
Google DeepMind achieves exceptional breakthrough in AI research
By
–
Truly exceptional achievement, bravo. @GoogleDeepMind researchers shine again 👏 https://t.co/Z6CGVJ0Rha
— Nando de Freitas (@NandoDF) 25 juillet 2024Truly exceptional achievement, bravo. @GoogleDeepMind researchers shine again
-
Meta Llama 3.1, OpenAI GPT-4o mini, and Apple AI models announced
By
–
@Meta dropped #Llama 3.1 405B, rivaling top models! OpenAI introduced GPT-4o mini, and @xai unveiled the Memphis Supercluster. @Apple
’s new AI models are now on @huggingface
. More details inside. #AI #Meta #OpenAI #Apple -
Foundation Model Research Sharing for Community Understanding
By
–
We hope that sharing this research will help the larger research community understand the key factors of foundation-model development and contribute to a more informed discussion about the future of foundation models in the general public.
-
Meta publishes comprehensive Llama 3 research paper
By
–
As part of our support of open science, we've published the full Llama 3 research paper covering a range of topics.
• Model training
• Model architecture
• Results of our ongoing work on integrating image/video/speech capabilities
• Much more Paper https://
go.fb.me/l5f55z -
Deep Learning PPG Smartphones Predict Cardiovascular Disease Early
By
–
Photoplethysmographs (PPGs) use light to measure variations in blood flow. Today we discuss new work that uses PPGs and deep learning to predict cardiovascular disease early, and release tooling for collecting PPG signals using smartphones. Learn more: https://t.co/KDPGvzIjPW pic.twitter.com/dbuLeK3pNW
— Google AI (@GoogleAI) 25 juillet 2024Photoplethysmographs (PPGs) use light to measure variations in blood flow. Today we discuss new work that uses PPGs and deep learning to predict cardiovascular disease early, and release tooling for collecting PPG signals using smartphones. Learn more: https://
goo.gle/4ffCGcr -
Jagged Intelligence: LLMs Excel and Fail Unpredictably
By
–
Jagged Intelligence The word I came up with to describe the (strange, unintuitive) fact that state of the art LLMs can both perform extremely impressive tasks (e.g. solve complex math problems) while simultaneously struggle with some very dumb problems. E.g. example from two