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

2020/11/14 by Bhaskar Mitra, Mitra, Bhaskar, Sebastian Hofstätter +5
Computer Science · #Topic Modeling #Natural Language Processing Techniques #Domain Adaptation and Few-Shot Learning

paper · pdf · doi:10.48550/arxiv.2011.07368

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.

Citations

• TU Wien @ TREC Deep Learning '19 -- Simple Contextualization for Re-ranking
• Overview of the TREC 2019 deep learning track
• Neural Methods for Effective, Efficient, and Exposure-Aware Information\n Retrieval
• Learning to Match Using Local and Distributed Representations of Text for Web Search

Related

• Conformer-Kernel with Query Term Independence for Document Retrieval
• Improving Transformer-Kernel Ranking Model Using Conformer and Query Term Independence
• Overview of the TREC 2019 deep learning track
• Overview of the TREC 2020 deep learning track
• Overview of the TREC 2021 deep learning track