Product Information
Description
Agent Memory and the Documentation Paradox: Knowledge Infrastructure for Long-Horizon Autonomous Systems — a survey of a frontier of autonomous AI in the Agents in the Wild series.
Book framework (chapter themes):
- Introduction — The Proposal of the Agent Memory and Documentation Paradox
- Agent Architecture and Cognitive Foundations: From the Perception–Action Loop to Memory Systems
- The Typology of Memory: Working Memory, Episodic Memory, and Procedural Memory
- Documentation as Memory Externalization: From Static Documents to Dynamic Knowledge Bases
- Memory Management for Long-Horizon Agents: Cross-Session State Retention and Knowledge Accumulation
- Memory-Augmented Agent Training: From Reinforcement Learning to Memory-Utility Optimization
- Memory Coordination and Knowledge Sharing in Multi-Agent Systems
- Document-Guided Agent Code Understanding and Migration
- Multimodal Memory — The Fusion of Visual, Textual, and Structured Evidence
- Memory Quality Evaluation — Benchmark Design Beyond Answer Correctness
- Memory and Documentation Interaction in Retrieval-Augmented Generation
- Memory Trust and Security — Poisoning, Misattribution, and Outdated Facts
- The Evolution and Reuse of Procedural Memory — Skill Acquisition and Interference
- Reasoning Enhancement of Agent Memory — Evidence Management in Long-Context Reasoning
- The Engineering of Memory Infrastructure — Storage, Retrieval, and Efficiency
- Empirical Research on the Documentation Paradox — Case Studies and Failure Modes
- Governance Framework for Agent Memory and Human Control
- Challenges, Limitations, and Open Questions
- Future Directions: Adaptive Memory, Neuro-Symbolic Fusion, and Cognitive Architecture
- Conclusion — Toward a Sustainable Agent Knowledge Infrastructure
Who it is for. Researchers with a foundation in machine learning and natural language processing, AI systems engineers, and advanced technical enthusiasts.
