Symbolic belief revision plus geometric projections lifts RAG reranking to 0.481 nDCG@10
letandrewcook · x · 2026-11-06
The author shares a retrieval-augmented method combining symbolic belief revision with geometric quality projections for stateful reranking.
Results:
- Full stateful reranker: 0.481 nDCG@10
- Graph-augmented baseline: 0.428
- Signals-only baseline: 0.400
That's +0.081 nDCG@10 over passage search and +0.053 over GraphRAG (p<0.001).
The core idea addresses a consistency problem in persistent interpretation: what happens to a derived proposition when its evidentiary basis changes. Each claim c is modeled with a set of alternative justifications J(c), support follows OR-of-AND semantics (supported(c) iff some justification has all antecedents supported), and the dependency graph must be acyclic.
Related event: Symbolic belief revision plus geometric projections boost RAG reranking(2 posts)→
More from coding & agent
- Click-to-summarize UI built on DocETL rewrite directives tames Anthropic's 421-page S-1 — sh_reya · 2026-11-07
- Amortized Reasoning Index: a retriever that self-improves during search, no labels needed — lateinteraction · 2026-11-07
- Linear token pricing breeds bloated agents — rationalized harnesses could run at ~300k tokens — JoshPurtell · 2026-11-07
- Linear token pricing kills incentives: optimized agents could run at 300k tokens — JoshPurtell · 2026-11-07
- Plug your phone into Codex: delete bloat apps, clear voicemail, tame old Kindles via ADB — angadsg · 2026-11-07
- 'Software factory' debate: marketing fluff or necessary infra for scaling coding agents? — vikvang1 · 2026-11-07