fleet-memory/hindsight-docs/docs/developer/retrieval.md
2025-11-27 16:22:16 +01:00

6.9 KiB

sidebar_position
3

Recall: How Hindsight Retrieves Memories

When you call recall(), Hindsight uses multiple search strategies in parallel to find the most relevant memories, regardless of how you phrase your query.

The Challenge of Memory Recall

Different queries need different search approaches:

  • "Alice works at Google" → needs exact name matching
  • "Where does Alice work?" → needs semantic understanding
  • "What did Alice do last spring?" → needs temporal reasoning
  • "Why did Alice leave?" → needs causal relationship tracing

No single search method handles all these well. Hindsight solves this with TEMPR — four complementary strategies that run in parallel.


Four Search Strategies

What it does: Understands the meaning behind words, not just the words themselves.

Best for:

  • Conceptual matches: "Alice's job" → "Alice works as a software engineer"
  • Paraphrasing: "Bob's expertise" → "Bob specializes in machine learning"
  • Synonyms: "meeting" matches "conference", "discussion", "gathering"

Why it matters: You can ask questions naturally without matching exact keywords.


What it does: Finds exact terms and names, even when they're spelled uniquely.

Best for:

  • Proper nouns: "Google", "Alice Chen", "MIT"
  • Technical terms: "PostgreSQL", "HNSW", "TensorFlow"
  • Unique identifiers: URLs, product names, specific phrases

Why it matters: Ensures you never miss results that mention specific names or terms, even if they're semantically distant from your query.


Graph Traversal

What it does: Follows connections between entities to find indirectly related information.

Best for:

  • Indirect relationships: "What does Alice do?" → Alice → Google → Google's products
  • Entity exploration: "Bob's colleagues" → Bob → co-workers → shared projects
  • Multi-hop reasoning: "Alice's team's achievements"

Why it matters: Retrieves facts that aren't semantically or lexically similar but are structurally connected through the knowledge graph.

Example: Even if Alice and her manager are never mentioned together, graph traversal can find the manager through shared projects or team relationships.


What it does: Understands time expressions and filters by when events occurred.

Best for:

  • Historical queries: "What did Alice do in 2023?"
  • Time ranges: "What happened last spring?"
  • Relative time: "What did Bob work on last year?"
  • Before/after: "What happened before Alice joined Google?"

How it works: Combines semantic understanding with time filtering to find events within specific periods.

Why it matters: Enables precise historical queries without losing old information.


Result Fusion

After the four strategies run, results are fused together:

  • Memories appearing in multiple strategies rank higher (consensus)
  • Rank matters more than score (robust across different scoring systems)
  • Final results are re-ranked using a neural model that considers query-memory interaction

Why fusion matters: A fact that's both semantically similar AND mentions the right entity will rank higher than one that's only semantically similar.


Token Budget Management

Results are automatically filtered to fit within your context window:

  • Top-ranked memories selected first
  • Stops when token budget is exhausted
  • Ensures you get the most relevant information within constraints

Parameters you control:

  • budget: Budget level for graph traversal (low=100, mid=300, high=600 nodes)
  • max_tokens: How much memory content to return (default: 4096 tokens)
  • fact_type: Filter by world, agent, opinion, or all

How Recall Works

When you call recall(query, bank_id):

  1. Parse → Detect temporal expressions, understand intent
  2. Search → Run 4 strategies in parallel
  3. Fuse → Combine results, prioritizing consensus
  4. Rerank → Neural reranking for final relevance
  5. Filter → Select top memories within token budget
  6. Return → Ranked, relevant memories

Tuning Recall: Quality vs Latency

Different use cases require different trade-offs between recall quality and response speed. Two parameters control this:

Budget: Graph Exploration Depth

Controls how many nodes to explore when traversing the knowledge graph:

Budget Nodes Explored Best For Trade-off
low 100 nodes Quick lookups, simple queries Fast, may miss distant connections
mid 300 nodes Most queries, balanced Good coverage, reasonable speed
high 600 nodes Complex multi-hop queries Thorough, slower

Example: "What did Alice's manager's team work on?" benefits from high budget to traverse Alice → manager → team → projects.

Max Tokens: Context Window Size

Controls how much memory content to return:

Max Tokens ~Pages of Text Best For Trade-off
2048 ~2 pages Focused answers, fast LLM Fewer memories, faster
4096 (default) ~4 pages Balanced context Good coverage, standard
8192 ~8 pages Comprehensive context More memories, slower LLM

Example: "Summarize everything about Alice" benefits from higher max_tokens to include more facts.

Trade-off Diagram

                Quality (Recall Coverage)
                        ↑
                        |
                  high budget
                  8192 tokens
                        |
            mid budget  |
            4096 tokens |
                        |
      low budget        |
      2048 tokens       |
                        |
        ────────────────┼────────────────→ Latency (Speed)
        Faster                    Slower
Use Case Budget Max Tokens Why
Chatbot replies low 2048 Fast responses, focused context
Document Q&A mid 4096 Balanced coverage and speed
Research queries high 8192 Comprehensive, multi-hop reasoning
Real-time search low 2048 Minimize latency

Why Multiple Strategies?

Consider the query: "What did Alice think about Python last spring?"

  • Semantic finds facts about Alice's opinions on programming
  • Keyword ensures "Python" is actually mentioned
  • Graph connects Alice → opinions → programming languages
  • Temporal filters to "last spring" timeframe

The fusion of all four gives you exactly what you're looking for, even though no single strategy would suffice.


Next Steps

  • Retain — How memories are stored with rich context
  • Reflect — How personality influences reasoning