Category: Machine Learning

• Meituan Quietly Launches New Model! Real-Test of First Open-Source "Heavy Thinking" Model: 8-Way Parallel, Agent Hard-Clashes with Claude
• Google's New Discovery: DeepSeek Reasoning Splits into Multiple Personalities, Left and Right Brain Competing for Intelligence
• Open-Source Framework Enables Code AI to Learn from GitHub! Bug Fix Rate Soars to 69.8%, Performance Sets New Records
• Google Just Overturned Model Memory, and Nvidia Revolutionized Attention｜Hao Good Chat Paper
• From 'LLM-as-a-Judge' to 'Agent-as-a-Judge': A Review of the Three-Stage Evolution of AI Evaluation Paradigms
• Optimization is Geometry, Geometry is Inference: Using Mathematics to End the Transformer Black Box Era
• Let AI Level Itself Up: Meta Pushes Coding to Superintelligence with Self-play RL
• Attention Is Not What You Need? Reframing Sequence Modeling with Geometric Aesthetics via Grassmann Manifolds
• From 'Titans+MIRAS & Nested' Architectural Innovations to NeurIPS2025 Best Paper 'Gated Attention'
• Under $8,000! Sina Weibo's 1.5B Small Model Surpasses Near-Trillion Parameter Models
• AI Cracks 18th-Century "Mystery" Ledger in Seconds! Google's New Model Blind Test Goes Viral
• SJTU PhD's Latest Insights: Clarifying Reinforcement Learning with Just Two Questions
• Meta Discovers: Slow RAG Systems Are Doing Too Much Unnecessary Work
• Recursive Reasoning HRM Model Reimagined! TRM Two-Layer Network (7M Parameters) Outperforms LLMs!
• Why Do Large Language Models Hallucinate? OpenAI's Latest Research Uncovers the Reasons
• DeepSeek, GPT-5's Fast-Slow Thinking Switching Gets a Smarter, Multimodal Version
• Stanford's Latest Research: Even the Strongest LLMs Struggle with Cutting-Edge Code! Gemini 2.5 Pro's Success Rate Under 40%
• LLMs Dominate Math Boards, Yet Forget How to Chat? CMU et al. Reveal Striking Differences Between SFT and RL!
• A New Revolution in Reward Models! SWIFT Reads "Inner Voice" Instead of Text, Creating a Faster, Stronger, and More Cost-Effective AI Judge
• Evolution and Development Trends of Reinforcement Learning Frameworks
• Advancing Silicon-Based Intelligence: Shuchao Bi's Insights on Past, Present, and Future AI
• AI's New SOTA in Bug Fixing: ExpeRepair Achieves 60.33% Fix Rate on SWE-Bench Lite, Learns from Experience Like Humans – Developed by Institute of Software, Chinese Academy of Sciences
• ReaGAN: Empowering Each Node as an Intelligent Reasoning Expert in Graphs
• Do Multimodal Large Language Models Truly 'Understand' the World? — Unveiling Core Knowledge Deficits in MLLMs
• How Mathematical Training "Unlocks" General Reasoning Abilities in Large Models? Latest Research Reveals Key Mechanisms