Leveraging Query Resolution and Reading Comprehension for Conversational Passage Retrieval

TL;DR

This paper presents a conversational passage retrieval system that combines BM25, BERT re-ranking, and query resolution with QuReTeC to improve retrieval accuracy in under-specified queries, demonstrating significant performance gains.

Contribution

It introduces a novel pipeline integrating query resolution with passage retrieval and re-ranking, enhancing performance in conversational information retrieval tasks.

Findings

Our system outperforms median runs in TREC CAsT 2020.

Query resolution with QuReTeC significantly improves retrieval accuracy.

Combining BM25, BERT, and comprehension models yields superior results.

Abstract

This paper describes the participation of UvA.ILPS group at the TREC CAsT 2020 track. Our passage retrieval pipeline consists of (i) an initial retrieval module that uses BM25, and (ii) a re-ranking module that combines the score of a BERT ranking model with the score of a machine comprehension model adjusted for passage retrieval. An important challenge in conversational passage retrieval is that queries are often under-specified. Thus, we perform query resolution, that is, add missing context from the conversation history to the current turn query using QuReTeC, a term classification query resolution model. We show that our best automatic and manual runs outperform the corresponding median runs by a large margin.