PLAID SHIRTTT for Large-Scale Streaming Dense Retrieval

arXiv:2405.00975v1 Announce Type: cross
Abstract: PLAID, an efficient implementation of the ColBERT late interaction bi-encoder using pretrained language models for ranking, consistently achieves state-of-the-art performance in monolingual, cross-language, and multilingual retrieval. PLAID differs from ColBERT by assigning terms to clusters and representing those terms as cluster centroids plus compressed residual vectors. While PLAID is effective in batch experiments, its performance degrades in streaming settings where documents arrive over time because representations of new tokens may be poorly modeled by the …

abstract art arxiv cluster cs.cl cs.ir encoder implementation language language models multilingual performance plaid ranking residual retrieval scale state streaming terms type vectors while