I agree with @ylecun that open source has been the primary reason for AI innovation and growth. Do not let misinformation and hysteria kill this. Open-source means we democratize and allow everyone to explore new ways to make AI reliable and safe, and allow for peer review.
@animaanandkumar
-
NeuralPlexer predicts protein-ligand dynamics with backbone changes
By
–
A great summary of our NeuralPlexer work which is the first to predict protein-ligand dynamics even when protein backbone may change during the process
-
AI Agents and Foundation Models Shape Future of Artificial Intelligence
By
–
A comprehensive report on state of #ai Happy to see our Voyager work featured in it. The future is not just foundation models like LLMs and diffusion models, but agents interacting with them in systematic ways to solve complex open-ended tasks. @yukez @DrJimFan @guanzhi_wang
-
GROOT Builds on MineDojo for Open-Ended Minecraft Learning
By
–
Great to see further work building on MineDojo. Minecraft is the right benchmark to test for open-ended learning. GROOT uses reference videos for instruction following and composes them together. https://t.co/cDgAzp3O0X
— Prof. Anima Anandkumar (@AnimaAnandkumar) 12 octobre 2023Great to see further work building on MineDojo. Minecraft is the right benchmark to test for open-ended learning. GROOT uses reference videos for instruction following and composes them together.
-
Fractional Fourier Neural Operator Advances Global Local Features
By
–
Nice work on Fractional Fourier neural operator. Fractional Fourier transform incorporates both time and frequency content that allows for both global and local features to be captured. Fun fact: my undergrad thesis used Fractional transform.
-
FourCastNet AI Model Outperforms Traditional Hurricane Prediction Methods
By
–
Great to see our FourCastNet model performing very well on hurricanes! "so far, the performance of AI models has been comparable to conventional models, sometimes better on tropical storm tracking. And the AI models do it fast, spitting out predictions on laptops within minutes"
-
FourCastNet AI Weather Model Forecasts Hurricane Lee Week Advance
By
–
Our #AI weather model FourCastNet nailed the forecast for Hurricane Lee a week in advance!
-
Neural Operators Enable Zero-Shot Super-Resolution with Sufficient Training Data
By
–
Indeed, neural operators can do zero-shot super-resolution as long as they are trained on data of sufficient resolution. There is a tradeoff between training costs and resulting fidelity. If we can get rich training data, then we can do faithful operator learning.
-
Training Data Resolution Insufficient for FNO Neural Operators
By
–
Thank you for creating this to publicize #AI for #PDE However, the training data is too low of a resolution and not enough to learn the operator faithfully for FNO and other neural operators. I also told @jo_brandstetter about this for PDE arena. This is not the right regime.
-
Integrating LLM inference in Lean via ONNX Runtime
By
–
Integrate Lean and LLM and an efficient way to run the inference in Lean. After training, the model is converted into the ONNX format, which can be run as a shared library using ONNX Runtime and integrated into Lean through FFI.