Conformer-Kernel with Query Term Independence at TREC 2020 Deep Learning Track

TL;DR

This paper evaluates Conformer-Kernel models on the TREC 2020 Deep Learning track, demonstrating that explicit term matching, query term independence, and ORCAS click data enhance retrieval performance.

Contribution

It introduces and benchmarks the impact of explicit term matching, query term independence, and ORCAS data within Conformer-Kernel models for information retrieval.

Findings

All three strategies improved retrieval quality.

Explicit term matching complements learned representations.

Query term independence enables scaling to full retrieval.

Abstract

We benchmark Conformer-Kernel models under the strict blind evaluation setting of the TREC 2020 Deep Learning track. In particular, we study the impact of incorporating: (i) Explicit term matching to complement matching based on learned representations (i.e., the "Duet principle"), (ii) query term independence (i.e., the "QTI assumption") to scale the model to the full retrieval setting, and (iii) the ORCAS click data as an additional document description field. We find evidence which supports that all three aforementioned strategies can lead to improved retrieval quality.