Overview of the TREC 2019 deep learning track

TL;DR

The TREC 2019 Deep Learning Track introduced large-scale datasets for ad hoc ranking, demonstrating that deep learning models significantly outperform traditional IR methods when trained on extensive human-labeled data.

Contribution

This paper presents the first large-scale deep learning track at TREC, with new datasets and evaluation methods for ad hoc ranking tasks.

Findings

Deep learning models outperformed traditional IR methods.

Large training datasets contributed to improved performance.

Deep models trained on extensive data showed significant advantages.

Abstract

The Deep Learning Track is a new track for TREC 2019, with the goal of studying ad hoc ranking in a large data regime. It is the first track with large human-labeled training sets, introducing two sets corresponding to two tasks, each with rigorous TREC-style blind evaluation and reusable test sets. The document retrieval task has a corpus of 3.2 million documents with 367 thousand training queries, for which we generate a reusable test set of 43 queries. The passage retrieval task has a corpus of 8.8 million passages with 503 thousand training queries, for which we generate a reusable test set of 43 queries. This year 15 groups submitted a total of 75 runs, using various combinations of deep learning, transfer learning and traditional IR ranking methods. Deep learning runs significantly outperformed traditional IR runs. Possible explanations for this result are that we introduced large training data and we included deep models trained on such data in our judging pools, whereas some past studies did not have such training data or pooling.