Hi,
I’m a postgraduate student in CS at Tianjin University.
Written for my present self, and also for someone in the future who might just need it.
This blog is intended to document my learning and research process, solutions to competition problems, and a few other interesting things.
Research areas currently under consideration:RSI, Agent, RL, Knowledge Enhancement
Writing
- AgentRx: Diagnosing AI Agent Failures from Execution TrajectoriesNov 11, 2026Reading notes on AgentRx, constraint validation, and root-cause diagnosis of agent failures.
- BLADE: Auditing Information Use for Failure AttributionNov 10, 2026Reading notes on BLADE, multi-agent failure attribution, and information-use auditing.
- DoVer: Intervention-Driven Multi-Agent DebuggingNov 6, 2026Reading notes on DoVer, multi-agent failure attribution, intervention, and validation.
- TrajDebug: Tracing Error LifecyclesNov 5, 2026Notes on TrajDebug, critical error localization, error lifecycle tracing, and feedback.
- Multi-Agent Failure Attribution Benchmark NotesNov 4, 2026Notes on the Who&When and MP-Bench multi-agent failure attribution benchmarks.
- Multi-Agent Systems Survey NotesOct 23, 2026Reading notes on Beyond Individual Intelligence: individual capabilities, reasoning, memory, planning, role assignment, multi-agent communication, orchestration, failure attribution, and self-evolution.
- Graph-based Agent Memory: Reading NotesOct 20, 2026The agent memory lifecycle and taxonomy, with graph-based memory storage, retrieval, and continual evolution.
- Dream-RSI: Reading NotesOct 16, 2026Notes on Dream-RSI: its motivation, online exploration and offline replay, and a few thoughts on its limitations.
- PPO to GRPOOct 11, 2026A complete derivation from policy gradient to GAE, importance sampling, and PPO-Clip, and GRPO
- KL DivergenceOct 9, 2026Notes on exact KL, k₁, k₃, and KL constraints and aggregation in GRPO.
- LoRAOct 8, 2026LoRA notes: low-rank updates, parameter counts, scaling, initialization, and common variants.
- MCP, A2A, ANPOct 5, 2026Notes on MCP, A2A, and ANP.
- Mini-SWE-Agent Source Code AnalysisSep 20, 2026A deep dive into the source code of Mini-SWE-Agent based on its execution trajectories, analyzing prompt injection, tool actions, exit mechanisms, and process management.
- AI Agent Learning Resources & RoadmapSep 16, 2026A curated collection of AI Agent learning resources and study paths gathered from various sources, to be continuously updated and expanded.
- 408 Exam Notes: Data Structures, Computer Organization, OS, and NetworksSep 9, 2026Complete revision notes for data structures, computer organization, operating-system scenarios, and computer networks.
- Graduate Entrance Exam Math Notes: Calculus and Linear AlgebraSep 9, 2026Complete revision notes for calculus and linear algebra.
- Daily Codeforces ProblemSep 7, 2026Starting Monday, solve one Codeforces problem each day, in order of increasing difficulty
- NOTE:Retrieval-Augmented GenerationAug 30, 2026Notes and Reflections on the Paper “Retrieval-Augmented Generation”
- Algorithm Notes: The Grand CollectionAug 29, 2026Two years of algorithm study notes from my cnblogs blog, reorganized into seven chapters: basic algorithms, search, dynamic programming, data structures, strings, graphs, and math
- Harness Engineering with HermesAug 8, 2026Building a full harness around an automated research-assistant scenario: plugin, permissions, auditing, three workflows, and automated tests