2026 · keynote
Beyond Hybrid Search with “Wormhole Vectors”
Wormhole Vectors are emerging as a technique to unify query understanding and retrieval across disparate vector spaces and query modalities (lexical, semantic, behavioral, etc.). This provides a significant improvement over typical hybrid search algorithms common in vector databases, search engines, RAG (retrieval augmented generation), and agentic search.
Most hybrid search implementations combine BM25 scoring on lexical keyword matches (best for matching specific words or labels) with semantic search on dense vector embeddings (better for matching the meaning of queries) and then fuse the results into a combined results list. Most agentic search approaches, likewise, treat lexical/BM25 and semantic/embedding search as independent tools that return separate sets of search results.
Wormhole vectors, in contrast, enable a unique form of pseudo-relevance feedback, utilizing the underlying documents in your search engine to find shared meaning(s) and query intention across these different vector spaces (sparse lexical space vs. dense semantic space). We can even build and leverage “behavioral” vector spaces based upon collaborative filtering, for introducing learned meaning from user interactions.
In this talk, we’ll show how hybrid search is typically implemented, how wormhole vectors work, and how to use them to traverse between disparate vector spaces as an improvement over hybrid search. You’ll see how to jump back and forth between multiple dense and sparse vector spaces in the same query, and we’ll show benchmarks and examples of how wormhole vectors compare to other leading query models and approaches like SPLADE and hybrid fusion algorithms. We’ll show with open source code and search engines how to implement these wormhole vectors, demonstrating their impact on search quality for traditional search, RAG, and agentic search.