Daily in-depth Chinese analyses of frontier AI research papers, including topics such as LLMs, agents, RAG, and RL. 每天深度解读AI相关论文

Reinforcement learning model for portfolio management that takes investor preferences into account

Accompanying code repository to the extended abstract: New Frontiers in Multi-Objective RL: Water Resource Management (AAMAS-2025)

Learning Efficient Frontier Selection with Cross-Attention Policy for Exploring Unstructured Environments [CORL 2025]

RL to discover prompt injection attacks in frontier LLMs

Formal comparison of SAR algorithm (frontier-based and RL-based)

Curated roadmap for World Models & Model-Based RL. Resources and projects to get hired at frontier AI and Robotics companies.

RL-style eval measuring intent/action divergence in frontier agents: model acknowledges a correction, then acts on the stale value anyway. 3 scenarios, 371 trials on claude-haiku-4-5, Sonnet 4.6, GPT-5.4, and Gemini 3.1 Pro Preview.

Optimizing Pareto frontiers in multi-objective RL environments

OpenEnv-compliant RL benchmark for measuring Agentic Betrayal in frontier LLMs

Write a simple agentic RL demo for learning and do experiment on some frontier concepts.

Execution-aware portfolio RL under transaction costs: target vs executed weights, cost alignment, η-frontier.

The Critical Path, Vol. I: RL at Scale — A Practical Guide to Reinforcement Learning for Frontier AI Systems

A two-layer dispatch optimizer that computes the exact Pareto frontier of cost-emission trade-offs over non-convex feasible sets, then tracks it in real time via entropy-regularized multi-agent RL under convex constraints.