DeepTileBars: Visualizing Term Distribution for Neural Information Retrieval
arXiv:1811.00606 · doi:10.1609/aaai.v33i01.3301289
Abstract
Most neural Information Retrieval (Neu-IR) models derive query-to-document ranking scores based on term-level matching. Inspired by TileBars, a classical term distribution visualization method, in this paper, we propose a novel Neu-IR model that handles query-to-document matching at the subtopic and higher levels. Our system first splits the documents into topical segments, "visualizes" the matchings between the query and the segments, and then feeds an interaction matrix into a Neu-IR model, DeepTileBars, to obtain the final ranking scores. DeepTileBars models the relevance signals occurring at different granularities in a document's topic hierarchy. It better captures the discourse structure of a document and thus the matching patterns. Although its design and implementation are light-weight, DeepTileBars outperforms other state-of-the-art Neu-IR models on benchmark datasets including the Text REtrieval Conference (TREC) 2010-2012 Web Tracks and LETOR 4.0.
References in corpus (12)
- Adam: A Method for Stochastic Optimization
- Distributed Representations of Words and Phrases and their Compositionality
- A Deep Relevance Matching Model for Ad-hoc Retrieval
- End-to-End Neural Ad-hoc Ranking with Kernel Pooling
- DeepRank: A New Deep Architecture for Relevance Ranking in Information Retrieval
- Introducing LETOR 4.0 Datasets
- Text Matching as Image Recognition
- Modeling Diverse Relevance Patterns in Ad-hoc Retrieval
- A Study of MatchPyramid Models on Ad-hoc Retrieval
- Learning Visual Features from Snapshots for Web Search
- A Scrollbar-based Visualization for Document Navigation
- Attention-based Neural Text Segmentation
Cited by in corpus (7)
- Information Retrieval: Recent Advances and Beyond
- Neural ranking models for document retrieval
- A Deep Look into Neural Ranking Models for Information Retrieval
- GazBy: Gaze-Based BERT Model to Incorporate Human Attention in Neural Information Retrieval
- Corpus-Level End-to-End Exploration for Interactive Systems
- Investigating Retrieval Method Selection with Axiomatic Features
- An In-depth Analysis of Passage-Level Label Transfer for Contextual Document Ranking