Top - machinelearning

Where do these stains come from and how can I fix them? (sopuli.xyz)

Hey guys,...

Where Is Noether's Principle in Machine Learning? | 2024-02-29 (cgad.ski)

2024-02-29 | Christopher Gadzinski writes:...

Gemini 1.5 (blog.google)

Anybody got to try it?

NumPy 2 is coming: preventing breakage, updating your code (pythonspeed.com)

Itamar Turner-Trauring writes:...

The 2023 Bletchley Declaration: A Major Leap in AI Safety and Ethics — Ultra Unlimited (www.ultra-unlimited.com)

Machine-learning system based on light could yield more powerful, efficient large language models (news.mit.edu)

I’ve had my eyes on optoelectronics as the future hardware foundation for ML compute (add not just interconnect) for a few years now, and it’s exciting to watch the leaps and bounds occurring at such a rapid pace.

Analysis suggests that no current AI systems are conscious, but also shows that there are no obvious barriers to building conscious AI systems. (arxiv.org)

RT-2: New model translates vision and language into action (www.deepmind.com)

[Paper] Learning to Generate Better Than Your LLM (arxiv.org)

I was looking through papers that combine LLMs and RL and this was pretty fascinating and the citations are perfect for continuing my search.

GitHub - aerdem4/lofo-importance: Leave One Feature Out Importance (github.com)

Almost All Research on the Mind is in English. That May Be a Problem (www.wired.com)

LongNet: Scaling Transformers to 1,000,000,000 Tokens (arxiv.org)

Stanford and Cornell Researchers Introduce Tart: An Innovative Plug-and-Play Transformer Module Enhancing AI Reasoning Capabilities in a Task-Agnostic Manner (www.marktechpost.com)

Meta AI -- I-JEPA: The first AI model based on Yann LeCun’s vision for more human-like AI (ai.facebook.com)

New model offers a way to speed up drug discovery (news.mit.edu)

Very cool application of LLM to screen potential protein-drug interactions.

JupyterLab 4.0 is Here (blog.jupyter.org)

Framework to play with statistical machine learning (www.cs.waikato.ac.nz)

Weka implements a ton of statistics-based ML algorithms as well as some validation tools and graphs. All you need is some data in almost-CSV format and you can run some statistics analysis on it. This isn’t neural networks so you don’t need neither a powerful GPU nor gigabytes of data. Some tutorials online get useful...