Apache Solr 10: What's Coming up for Vector Search
summary
Solr 10/10.1 brings concrete vector-search plumbing: scalar quantization (compress float32 to 4 or 7 bits) and binary quantization (index as 1‑bit vectors, query encoded into 4‑bit) reduce memory and speed queries at measurable accuracy cost; HNSW gains early termination that exits when the nearest‑neighbor set stabilizes (default saturation 0.995, patience = max(30% of topK, 7)); seeded KNN starts graph walks from a lexical seed, ACORN‑style filtered traversal triggers below ~60% matches, and an EF search scale factor expands ef_search — all implemented in Lucene and exposed via Solr’s KNN query parser.
AI-written summary · about
chapters (9)
- 0:00Introduction and speakers
- 1:33Apache Solr 10 vector search features
- 3:42Quantization for vector search
- 12:36Early termination in HNSW search
- 20:16Seeded KNN vector query
- 24:17Filtered search threshold with Acrony
- 29:18EF search scale factor
- 32:52Upcoming features in Solr 10.1
- 37:43Q&A session
related talks
28mHybrid Search with Apache Solr
47mNeural Search Comes to Apache Solr: Approximate Nearest Neighbor, BERT & more
44mVector Search Beyond the Hype: Real-World Implementation and Performance Insights
43mOpenSearch and the Path to a Trillion-scale Vector Database
01h 09mVector Search: Ask Me Anything!
37mLexical, Vector & Hybrid Search with Elasticsearch
46mHibernate Search 7: What's New and Practical Cases for Java Full-Text and Vector Search
24mVector Search at Uber