Category: Machine Learning

• Train a Tiny LLM from Scratch for Just ¥8 in 9 Hours! Full Tutorial Including Reasoning, MoE, and More
• More Capable Than Gemini Diffusion! The First Multimodal Large Diffusion Language Model MMaDA Released, Achieving Strong Reasoning and High Controllability
• OpenAI's Big Move! Core API Now Supports MCP, Revolutionizing Agent Development Overnight
• Does AI Know When to "Think"? Thinkless Teaches Large Language Models When to Reason
• ICML 2025 | Training-Free, Instant Alignment of Large Model Preferences
• First Author Interpretation! Talking About Qwen's New Scaling Law—Parallel Scaling—From an Idea Perspective
• Breakthrough in Reasoning: How SoftCoT++ Enables LLMs to 'Think Multiple Paths'?
• Why We’re Unlikely to Get Artificial General Intelligence Anytime Soon
• ZeroSearch: <Alibaba Technology> Large Language Models Learn Through Self-Rewarding Without a Browser
• 10 Years of Hard Research, Millions Wasted! AI Black Box Remains Unsolvable, Google Breaks Face-off
• Continuous Thought Machines Are Here! Startup by a 'Transformer Eight Sons' Member Launches, Letting AI Stop Making 'One-Step' Snap Decisions
• Stanford's Weak-for-Strong (W4S): Harnessing Stronger LLMs with Meta-Agent, Accuracy Boosted to 95.4% | Latest
• The 'era of experience' will unleash self-learning AI agents across the web—here's how to prepare
• ByteDance Seed's New Method! Open-Source 8B Code Model: Trains Itself by Curating Its Own Data, Achieves SoTA at Its Scale, and Even Surpasses 10 Billion Parameter Competitors
• "Absolute Zero": A Zero-Data, Self-Evolving AI Reasoning Method Surpasses SOTA
• PKU, Tsinghua, UvA, CMU, etc. Jointly Release: Latest Survey on Logical Reasoning Abilities of Large Models
• A Self-Improving Coding Agent
• Bridging the Gap: LUFFY, a New Reinforcement Learning Paradigm for AI Reasoning
• First Chapter of 'Reasoning From Scratch' Released: Sebastian Raschka on LLM Reasoning, Pattern Matching, and Foundational Training
• DeepSeek makes a big move! New model focuses on mathematical theorem proving, significantly refreshing multiple high-difficulty benchmarks.