# A Reproducibility Study of PLAID

arXiv: https://arxiv.org/abs/2404.14989 · HTML: https://mlsanity.com/paper/2404.14989/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 28 references that are in the map.
- First posted on arXiv 23 Apr 2024.

## Builds on (most cited first, 15 of 15)

- [Dense Passage Retrieval for Open-Domain Question Answering](https://mlsanity.com/paper/2004.04906/) (2020)
- [MS MARCO: A Human Generated MAchine Reading COmprehension Dataset](https://mlsanity.com/paper/1611.09268/) (2016)
- [ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT](https://mlsanity.com/paper/2004.12832/) (2020)
- [ColBERTv2: Effective and Efficient Retrieval via Lightweight Late Interaction](https://mlsanity.com/paper/2112.01488/) (2021)
- [Overview of the TREC 2019 deep learning track](https://mlsanity.com/paper/2003.07820/) (2020)
- [PLAID: An Efficient Engine for Late Interaction Retrieval](https://mlsanity.com/paper/2205.09707/) (2022)
- [CEDR: Contextualized Embeddings for Document Ranking](https://mlsanity.com/paper/1904.07094/) (2019)
- [Pytrec_eval: An Extremely Fast Python Interface to trec_eval](https://mlsanity.com/paper/1805.01597/) (2018)
- [Pseudo-Relevance Feedback for Multiple Representation Dense Retrieval](https://mlsanity.com/paper/2106.11251/) (2021)
- [MS MARCO: Benchmarking Ranking Models in the Large-Data Regime](https://mlsanity.com/paper/2105.04021/) (2021)
- [Interpretable & Time-Budget-Constrained Contextualization for Re-Ranking](https://mlsanity.com/paper/2002.01854/) (2020)
- [A Unified Framework for Learned Sparse Retrieval](https://mlsanity.com/paper/2303.13416/) (2023)
- [Lexically-Accelerated Dense Retrieval](https://mlsanity.com/paper/2307.16779/) (2023)
- [How to Measure the Reproducibility of System-oriented IR Experiments](https://mlsanity.com/paper/2010.13447/) (2020)
- [The tale of two MS MARCO -- and their unfair comparisons](https://mlsanity.com/paper/2304.12904/) (2023)

## Built on by (most cited first, 8 of 8)

- [Guiding Retrieval using LLM-based Listwise Rankers](https://mlsanity.com/paper/2501.09186/) (2025)
- [JaColBERTv2.5: Optimising Multi-Vector Retrievers to Create State-of-the-Art Japanese Retrievers with Constrained Resources](https://mlsanity.com/paper/2407.20750/) (2024)
- [Efficient Constant-Space Multi-Vector Retrieval](https://mlsanity.com/paper/2504.01818/) (2025)
- [Simple Projection Variants Improve ColBERT Performance](https://mlsanity.com/paper/2510.12327/) (2025)
- [LEMUR: Learned Multi-Vector Retrieval](https://mlsanity.com/paper/2601.21853/) (2026)
- [Efficient Multivector Retrieval with Token-Aware Clustering and Hierarchical Indexing](https://mlsanity.com/paper/2604.28142/) (2026)
- [ColBERTSaR: Sparsified ColBERT Index via Product Quantization](https://mlsanity.com/paper/2606.05568/) (2026)
- [PLAID-PRF: Pseudo-Relevance Feedback with Centroid-like Tokens in PLAID](https://mlsanity.com/paper/2607.18626/) (2026)

## Abstract (arXiv)

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. By studying the parameters PLAID introduces, we find that its Pareto frontier is formed of a careful balance among its three 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 provides better efficiency-effectiveness trade-offs in low-latency settings. 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. 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 when evaluated using a well-annotated dataset. 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 the competitive trade-offs that re-ranking baselines exhibit, this work highlights the importance of carefully selecting pertinent baselines when evaluating the efficiency of retrieval engines.