Content-Based Weak Supervision for Ad-Hoc Re-Ranking

Abstract

One challenge with neural ranking is the need for a large amount of manually-labeled relevance judgments for training. In contrast with prior work, we examine the use of weak supervision sources for training that yield pseudo query-document pairs that already exhibit relevance (e.g., newswire headline-content pairs and encyclopedic heading-paragraph pairs). We also propose filtering techniques to eliminate training samples that are too far out of domain using two techniques: a heuristic-based approach and novel supervised filter that re-purposes a neural ranker. Using several leading neural ranking architectures and multiple weak supervision datasets, we show that these sources of training pairs are effective on their own (outperforming prior weak supervision techniques), and that filtering can further improve performance.

arXiv commentSIGIR 2019 (short paper)

References in corpus (5)

• A Deep Relevance Matching Model for Ad-hoc Retrieval
• End-to-End Neural Ad-hoc Ranking with Kernel Pooling
• IRGAN: A Minimax Game for Unifying Generative and Discriminative Information Retrieval Models
• A Study of MatchPyramid Models on Ad-hoc Retrieval
• Overcoming low-utility facets for complex answer retrieval

Cited by in corpus (8)

• Simplified Data Wrangling with ir_datasets
• Efficient Document Re-Ranking for Transformers by Precomputing Term Representations
• PROP: Pre-training with Representative Words Prediction for Ad-hoc Retrieval
• ABNIRML: Analyzing the Behavior of Neural IR Models
• Learnable Pillar-based Re-ranking for Image-Text Retrieval
• Selective Weak Supervision for Neural Information Retrieval
• Few-Shot Text Ranking with Meta Adapted Synthetic Weak Supervision
• Towards a Relevance Posterior in Neural Information Access