PathSeer: Adaptive Filtered Vector Search

Filtered vector search combines semantic similarity with structured conditions, and Weaviate 1.40 introduces PathSeer, a predicate-agnostic HNSW filtering strategy based on SIGMOD 2026 research that dynamically balances distance-first traversal, which preserves graph connectivity but can waste computations, with filter-first traversal, which reduces unnecessary distance calculations but can harm recall. Weaviate’s adaptation works on existing HNSW graphs without rebuilding indexes, initially using a limited two-hop expansion to find matching candidates and then retaining nonmatching nodes as potential bridges while avoiding some distance calculations. Benchmarks across BEIR, Sphere DPR, and 10-million-item wiki-dpr datasets found that PathSeer generally matched or exceeded sweeping and ACORN for topical, random, and conjunction filters, particularly improving throughput at a 0.90 recall target, although ACORN remained substantially faster for strongly negatively correlated filters. PathSeer can be enabled through the mutable HNSW `filterStrategy` setting without reindexing, while Weaviate’s existing brute-force cutoff continues to handle filters with very small result sets.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.