sqpe.devContribute
← Back to research
2026 / Optimization

Fast semantic filtering

Fast LLM-Based Semantic Filtering: From a Unified Framework to an Adaptive Two-Phase Method

What it contributes

Combines clustering and learned proxies adaptively, using token-aware representations and oracle confidence to reduce expensive filtering calls.

How to read the evidence

Empirical attainment of an accuracy target is not an unconditional guarantee; the oracle itself may be wrong.

We checked the primary publication record and abstract. This profile summarizes author-reported work; it does not claim a full-paper audit, implementation audit, or independent reproduction.

Explore the connections

FiltersCascadesProxy models
  • LOTUS — Relational operators
  • JEVDB — SQL / typed decisions
  • Larch — AI SQL filters