Agent Memory: An Anatomy
The article discusses the structure and function of agent memory libraries, emphasizing the distinction between different types of memory. It highlights how these libraries often use cognitive science terminology but may not fully embody the complexities of human memory. Key components of agent memory systems include extractors, stores, and retrievers, each playing a crucial role in how information is processed and retrieved.
- ▪Agent memory libraries utilize terms like episodic and semantic memory from cognitive science.
- ▪The extractor component determines what information to keep from conversations, impacting the context available for retrieval.
- ▪The store component manages how statements are stored and how to handle contradictions in user information.
Opening excerpt (first ~120 words) tap to expand
[ ← INDEX ] agent memory: an anatomy 26 MAY 2026 every agent memory library uses the same words: episodic, semantic, sometimes procedural. they’re cognitive science’s vocabulary, lifted into the API. the engineering often isn’t lifted with them. a library can have a procedural field that uses the same storage and retrieval as semantic — a label, not a separate system. the deeper slip is the word memory itself: most of what these libraries build is narrower than that, and the narrower term sharpens the problem. the terminology comes from a 1972 chapter by Endel Tulving.1 he argued that what people had been treating as one thing — memory — was at least two: memory for events (what happened, where, when), and memory for facts (the capital of France, water’s boiling point).
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at brgsk.