Source: Santhanam et al. 2022: PLAID
Paper describing PLAID, the ColBERTv2 search engine that prunes candidates using centroids, cutting latency up to 7x on GPU and 45x on CPU.
Contents
- What: research paper, arXiv:2205.09707
- Author: Keshav Santhanam, Omar Khattab, Christopher Potts, Matei Zaharia
- Date: 2022-05-19
- Original: PLAID: An Efficient Engine for Late Interaction Retrieval
Summary
Exploits the centroids already present in ColBERTv2’s compressed index: each passage is first treated as a cheap “bag of centroids”, scored against the query, and pruned, so the expensive residual decompression and full MaxSim run only on a small survivor set.
Key claims
- Centroid interaction quickly discards low-scoring passages; centroid pruning sparsifies each passage’s bag of centroids.
- Up to 7x lower latency on GPU and 45x on CPU versus vanilla ColBERTv2, without quality loss.
- Tens of milliseconds on GPU and tens to a few hundred milliseconds on CPU at 140M passages.