Retrieval with Multiple Query Vectors through Anomalous Pattern Detection
For practitioners needing complex retrieval with multiple query vectors, this work offers a novel approach, though improvements are incremental and dataset-dependent.
The paper proposes a retrieval method that handles multiple query vectors simultaneously by identifying anomalous dimensions in the query set and retrieving database vectors with similar anomalies. Experiments on four datasets show that larger query sets improve retrieval performance, with the most significant gains from 1 to 8 queries.
A classical vector retrieval problem typically considers a \emph{single} query embedding vector as input and retrieves the most similar embedding vectors from a vector database. However, complex reasoning and retrieval tasks frequently require \emph{multiple query vectors}, rather than a single one. In this work, we propose a retrieval method that considers multiple query vectors simultaneously and retrieves the most relevant vectors from the database using concepts from anomalous pattern detection. Specifically, our approach leverages a set of query vectors $Q$ (with $|Q|\geq 1$), and identifies the subset of vector dimensions within $Q$ that standout (anomalous) from the rest of dimensions. Next, we scan the vector database to retrieve the set of vectors that are also anomalous across the previously identified vector dimensions and return them as our retrieved set of vectors. We validate our approach on two image datasets, a text dataset, and a tabular dataset. Overall, we observe that, across most datasets, larger query sets lead to improved retrieval performance. The improvement is most pronounced when increasing the query sets from 1 to 8, while the gains become smaller beyond that.