Generative Retrieval for Book search

2025-01-19

AI Summary

To address challenges in book search—including difficulty integrating multi-source heterogeneous information (e.g., metadata, table of contents, and full text), weak hierarchical structure modeling, and scarcity of labeled training data—this paper proposes the Generative Book Search (GBS) framework. GBS introduces two key innovations: (1) a table-of-contents-guided book encoding mechanism that explicitly models chapter-level hierarchical relationships via dual-layer positional encoding and preservation-aware attention; and (2) a pseudo-query generation strategy that jointly optimizes coverage and diversity, leveraging the table of contents to construct both pseudo-queries and book identifiers—thereby mitigating label scarcity. Evaluated on Baidu’s internal book dataset, GBS achieves a 9.8% improvement in MRR@20 over RIPOR, demonstrating substantial gains in long-context understanding and structural awareness.

Technology Category

• Search and Optimization: Learning to Search
• Natural Language Processing: Learning & Optimization for NLP
• Data Mining & Knowledge Management: Intelligent Query Processing

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs
• Semantics and Knowledge: Applications of semantic technologies for improving search, browsing, recommendation, personalization

Abstract

In book search, relevant book information should be returned in response to a query. Books contain complex, multi-faceted information such as metadata, outlines, and main text, where the outline provides hierarchical information between chapters and sections. Generative retrieval (GR) is a new retrieval paradigm that consolidates corpus information into a single model to generate identifiers of documents that are relevant to a given query. How can GR be applied to book search? Directly applying GR to book search is a challenge due to the unique characteristics of book search: The model needs to retain the complex, multi-faceted information of the book, which increases the demand for labeled data. Splitting book information and treating it as a collection of separate segments for learning might result in a loss of hierarchical information. We propose an effective Generative retrieval framework for Book Search (GBS) that features two main components: data augmentation and outline-oriented book encoding. For data augmentation, GBS constructs multiple query-book pairs for training; it constructs multiple book identifiers based on the outline, various forms of book contents, and simulates real book retrieval scenarios with varied pseudo-queries. This includes coverage-promoting book identifier augmentation, allowing the model to learn to index effectively, and diversity-enhanced query augmentation, allowing the model to learn to retrieve effectively. Outline-oriented book encoding improves length extrapolation through bi-level positional encoding and retentive attention mechanisms to maintain context over long sequences. Experiments on a proprietary Baidu dataset demonstrate that GBS outperforms strong baselines, achieving a 9.8% improvement in terms of MRR@20, over the state-of-the-art RIPOR method...

Problem

• Generative Retrieval
• Book Recommendation
• Unlabeled Data

Innovation

• GBS Framework
• Data Augmentation
• Table-of-Contents Encoding