I am thrilled to see @GoogleDeepMind build on our work in several instances to strengthen contributions to AI+Science and Math: 1. More than a year ago, we released the first open-source framework, https://
leandojo.org, using Lean+LLM for theorem proving and an AI Lean
@animaanandkumar
-
DeepMind Builds on LeanDojo for AI-Powered Theorem Proving
By
–
-
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
… -
Long-Context Transformers: Training Llama3 with 60k Token Sequences
By
–
Introducing long-context transformer using mini sequences. It is a simple and effective method for highly efficient and accurate LLM training with extremely long sequences. Our research demonstrates that the Llama3-8B model can be trained with context lengths up to 60k tokens on
-
GPT-2 Sized Model Demonstrates Universal Physical Understanding
By
–
Building AI with universal physical understanding will have a tremendous impact. We have taken the first steps towards this by building a ~GPT-2 sized model that shows that a broad model can capture a wide range of physical phenomena. https://
arxiv.org/pdf/2403.03542 We have scaled the -
FourCastNet AI Model Outperforms Traditional Weather Forecasting Systems
By
–
Thrilled to see that our AI-based weather model FourCastNet is again beating traditional forecasting systems on #HurricaneBeryl We saw this also with #hurricane Harvey last year. FourCastNet was the first AI-based weather model. We trained it on historical weather data
-
IEEE Kiyo Tomiyasu Award for Innovative Technology Contributions
By
–
I am thrilled to receive the @IEEEorg Kiyo Tomiyasu Award for outstanding early to mid-career contributions to technologies holding the promise of innovative applications.
— Prof. Anima Anandkumar (@AnimaAnandkumar) 6 juillet 2024
I had no expectations of getting such a broad award that spans all areas of IEEE. It is an honor to be in… https://t.co/HvgyAPLlhHI am thrilled to receive the @IEEEorg Kiyo Tomiyasu Award for outstanding early to mid-career contributions to technologies holding the promise of innovative applications. I had no expectations of getting such a broad award that spans all areas of IEEE. It is an honor to be in
-
Neural Operators: Advanced Machine Learning Research Overview
By
–
Check out Neural Operators if you haven't already done so! https://
x.com/AnimaAnandkuma
r/status/1789384583563431937
… -
Distinguished Alumni Award at IIT Madras After Twenty Years
By
–
Looking forward to being back at @iitmadras It has been 20 years since I left but memories are still fresh. It is an honor to be back to receive the distinguished alumni award.
-
Fourier Neural Operators: Architecture, Theory, and Field Overview
By
–
The original FNO paper gives one specific architecture, the JMLR paper provides more of a theoretical foundation, and the nature perspective article gives a more general view of the field. https://
rdcu.be/dD6LO -
Neural Operators Enable Resolution-Agnostic Function Space Mappings
By
–
Neural operators are an approach to generalize to new resolutions since they learn mappings between function spaces.