How RecursiveMAS speeds up multi-agent inference by 2.4x and reduces token usage by 75%

UIUC and Stanford's RecursiveMAS lets AI agents collaborate in embedding space instead of text, cutting token usage by 75% ...

AI generates first complete models of proteins in motion

Many drug and antibody discovery pathways focus on intricately folded cell membrane proteins. When molecules of a drug ...
marktechpost

Qwen AI Releases Qwen-Scope: An Open-Source Sparse AutoEncoders (SAE) Suite That Turns LLM Internal Features into Practical Development Tools

Large language models are remarkably capable, yet frustratingly opaque. When a model misbehaves — generating responses in the wrong language, repeating itself endlessly, or refusing safe requests — AI ...
GitHub

SARMAE: Masked Autoencoder for SAR Representation Learning

Synthetic Aperture Radar (SAR) imagery plays a critical role in all-weather, day-and-night remote sensing applications. However, existing SAR-oriented deep learning is constrained by data scarcity, ...
The New York Times

A.I. Replica of Val Kilmer to Appear in Film After His Death

Kilmer’s estate collaborated with the filmmakers to make a digital likeness of the actor, who died of cancer last year. By Reggie Ugwu A photorealistic A.I. likeness of Val Kilmer, the “Top Gun,” ...
IEEE

Explainable Deep Learning for Cyber Threat in IoMT: A Synchronization-Enhanced Sparse Autoencoder Approach

Abstract: The interconnected medical devices and networks face escalating cyber threats that demand intelligent detection mechanisms. Traditional deep learning approaches for cyber threat intelligence ...
Frontiers

Adaptive multi-mode locomotion for bipedal wheel-legged robots via sparse mixture-of-experts deep reinforcement learning

The bipedal wheel-legged robot combines the high energy efficiency of wheeled movement with the terrain adaptability of legged locomotion. However, achieving a smooth transition between these two ...
IEEE

Sparse Autoencoder-Based Deep Learning Approach for Enhancing End-to-End Communication System Performance

Deep Learning (DL) has emerged as a transformative approach in artificial intelligence, demonstrating remarkable capabilities in solving complex problems once considered unattainable. Its ability to ...
GitHub

UNITES-Lab/sparse-cafm

@inproceedings{10.1117/12.3067427, author = {Levi Harris and Md Jayed Hossain and Mufan Qui and Ruichen Zhang and Pingchuan Ma and Tianlong Chen and Jiaqi Gu and Seth ...
blockchain

Understanding Neural Networks Through Sparse Circuits: OpenAI's Breakthrough in Interpretable AI Models

According to Sam Altman on Twitter, OpenAI has shared insights on understanding neural networks through sparse circuits, offering a practical approach to improve ...