Most agent memory implementation rely on basic vector search or sometimes use a knowledge graph. Hindsight uses biomimetic data structures to organize agent memories in a way that is more like how human memory works:
- World: Facts about the world ("The stove gets hot")
- Experiences: Agent's own experiences ("I touched the stove and it really hurt")
- Opinion: Beliefs with confidence scores ("I shouldn't touch the stove again" - .99 confidence)
- Observation: Complex mental models derived by reflecting on facts and experiences ("Curling irons, ovens, and fire are also hot. I shouldn't touch those either.")
Memories in Hindsight are stored in banks (i.e. memory banks). When memories are added to Hindsight, they are pushed into either the world facts or experiences memory pathway. They are then represented as a combination of entities, relationships, and time series with sparse/dense vector representations to aid in later recall.
Hindsight provides three simple methods to interact with the system:
- Retain: Provide information to Hindsight that you want it to remember
- Recall: Retrieve memories from Hindsight
- Reflect: Reflect on memories and experiences to generate new observations and insights from existing memories.
Agent Memory That Learns
A key goal of Hindsight is to build agent memory that enables agents to learn and improve over time. This is the role of the reflect operation which provides the agent to form broader opinions and observations over time.
For example, imagine a product support agent that is helping a user troubleshoot a problem. It uses a search-documentation t
|