Toyota Research Unveils Driving Foundation Model
Toyota Research Institute announces a new foundation model designed to tackle complex urban autonomous driving scenarios…
📂 Research
Latest articles in Research
Berkeley AI Lab Releases Open Source RLHF Framework
UC Berkeley's AI research lab publishes a comprehensive open source framework for RLHF training of large language models…
New Architecture Rivals Transformers in AI
Researchers unveil a novel architecture that challenges the dominance of attention-based Transformers with better effici…
Google Scales Sparse MoE Models to Trillion Params
Google Research introduces Sparse Mixture of Experts architecture that scales language models to over 1 trillion paramet…
CMU Researchers Crack Multimodal Medical AI Diagnosis
Carnegie Mellon team develops a multimodal AI system that integrates imaging, text, and lab data to diagnose diseases wi…
MIT CSAIL Unveils RL Breakthrough for Robot Dexterity
MIT CSAIL researchers introduce a novel reinforcement learning framework that dramatically improves robotic manipulation…
SubQ: Sub-Quadratic LLM Handles 12M-Token Context
SubQ introduces a sub-quadratic architecture enabling LLMs to process up to 12 million tokens, shattering previous conte…
Google Brain Unveils Mixture-of-Depths Architecture
Google Brain's new Mixture-of-Depths transformer architecture dynamically allocates compute per token, cutting inference…
Anthropic Reveals How Claude 4 Actually Thinks
Anthropic publishes landmark mechanistic interpretability research mapping internal reasoning circuits in Claude 4 model…
NVIDIA Unveils Diffusion Transformer for Real-Time 3D
NVIDIA publishes a novel Diffusion Transformer architecture that enables real-time 3D asset generation, potentially tran…
UC Berkeley AI Agent Masters Robotic Surgery Tasks
UC Berkeley researchers develop a reinforcement learning agent capable of performing complex robotic surgery tasks with …
Meta FAIR Launches V-JEPA 2.0 Video AI Model
Meta's FAIR lab unveils V-JEPA 2.0, a self-supervised model that understands video without labeled data, marking a major…