Best Paper Runner Up pdf bibtex slides 2 citations reproducibility paper
Appeared in: Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR 2024)
Abstract:
The PLAID (Performance-optimized Late Interaction Driver) algorithm for ColBERTv2 uses clustered term representations to retrieve and progressively prune documents for final (exact) document scoring. In this paper, we reproduce and fill in missing gaps from the original work. First, by studying the parameters PLAID introduces, we find that its Pareto frontier is formed of a careful balance among its parameters; deviations beyond the suggested settings can substantially increase latency without necessarily improving its effectiveness. We then compare PLAID with an important baseline missing from the paper: re-ranking a lexical system. We find that applying ColBERTv2 as a re-ranker atop an initial pool of BM25 results consistently provides a better efficiency-effectiveness trade-off. However, re-ranking cannot reach peak effectiveness at higher latency settings due to limitations in recall of lexical matching and provides a poor approximation of an exhaustive ColBERTv2 search. However, we find that recently proposed modifications to re-ranking that pull in the neighbors of top-scoring documents overcome this limitation, providing a Pareto frontier across all operational points for ColBERTv2. Curious about why re-ranking methods are highly competitive with PLAID, we analyze the token representation clusters PLAID uses for retrieval and find that most clusters are predominantly aligned with a single token and vice versa. Given that re-ranking is considerably simpler than PLAID, this work highlights the importance of carefully selecting pertinent baselines when evaluating the efficiency of retrieval engines.
BibTeX @inproceedings{macavaney:sigir2024-plaidrepro, author = {MacAvaney, Sean and Tonellotto, Nicola}, title = {A Reproducibility Study of PLAID}, booktitle = {Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval}, year = {2024}, url = {https://arxiv.org/abs/2404.14989}, doi = {10.1145/3626772.3657856} }