RESEARCH - AI Dynamics

AI Physicists: From Computational Simulations to Autonomous Law Discovery

By

–

07 July 2023 18h22

on physics and AI— today, all theoretical physicists are computational—they code simulations & draw theory from them tmrw, all theoretical physicists will be AI physicists—AI will derive laws of physics from observational data, and humans will filter them to find the good laws

→ View original post on X — @alexandr_wang

7 July 2023

MIT Team Advances 3D Packing Using Fast Fourier Transform

By

@mit_csail

–

07 July 2023 18h00

NP-hard has long complicated 3D packing, but an MIT-led team applied the fast Fourier transform to the problem.

The result: a novel packing algorithm that searches for placement locations in an object, potentially making 3D packing much more efficient: https://t.co/BtzftXNP7N pic.twitter.com/40hpupKnkn
— MIT CSAIL (@MIT_CSAIL) 7 juillet 2023

NP-hard has long complicated 3D packing, but an MIT-led team applied the fast Fourier transform to the problem. The result: a novel packing algorithm that searches for placement locations in an object, potentially making 3D packing much more efficient: http://
bit.ly/3NIRh32

→ View original post on X — @mit_csail

7 July 2023

Can You Trust ChatGPT’s Package Recommendations?

By

@miketamir

–

07 July 2023 17h37

Can you trust ChatGPT’s package recommendations? https://
bit.ly/3CAELxE #AI #MachineLearning #DeepLearning #LLMs #DataScience

→ View original post on X — @miketamir

7 July 2023

BigBird Encoder vs Decoder LLM Architecture Comparison

By

@rasbt

–

07 July 2023 17h11

It's been a while and would have to check again, but wasn't BigBird a encoder LLM? Versus this is a decoder-style LLM for generating texts. But generally yes!

→ View original post on X — @rasbt

7 July 2023

LongNet: Scaling Transformers to 1 Billion Tokens with Dilated Attention

By

@rasbt

–

07 July 2023 17h02

We had 1) the RMT paper on scaling Transformers to 1M tokens, and 2) the convolutional Hyena LLM for 1M tokens. How about 3) LONGNET: Scaling Transformers to 1 Billion Tokens (
https://
arxiv.org/abs/2307.02486)!? It achieves linear (vs quadratic) scaling via dilated (vs self) attention.

→ View original post on X — @rasbt

7 July 2023

Vector Quantised TTS Model Training Collaboration Sought

By

@reach_vb

–

07 July 2023 16h31

Who out here is training a vector quantised TTS model? Looking for collaborations to bring easy-to-scale training recipes for TTS.
DM?

→ View original post on X — @reach_vb

7 July 2023

AlphaFold Wins Gairdner Award Alongside John Jumper

By

@demishassabis

–

07 July 2023 15h51

Hugely honoured to be recognised by the @GairdnerAwards alongside John Jumper and on behalf of the entire @DeepMind team for our work in creating #AlphaFold and the impact it has had

→ View original post on X — @demishassabis

7 July 2023