Neural Approaches to Conversational Information Retrieval

Recent developments in Conversational Information Retrieval (CIR) are surveyed, with a focus on neural approaches that have emerged in recent years. The book discusses techniques for evaluating CIR systems, algorithms and methods for developing main modules, and models for conversational machine comprehension and question answering. It also reviews commercial systems for CIR and related tasks. The target audience includes IR and NLP research communities, as well as those with backgrounds in machine learning or human-computer interaction.

Format: Hardback
Length: 211 pages
Publication date: 18 March 2023
Publisher: Springer International Publishing AG

The field of Conversational Information Retrieval (CIR) has witnessed significant advancements in recent years, driven by the remarkable progress in deep learning. This comprehensive book delves into the latest developments in neural approaches, highlighting their profound impact on natural language processing (NLP) and conversational AI. The result is a multitude of commercial conversational services that facilitate seamless and intuitive interactions, fostering a more human-centric approach to information retrieval.

Comprising nine chapters, this book offers a comprehensive exploration of CIR. Chapter 1 sets the stage by reviewing the studies on human search behavior and defining a CIR system, along with a detailed reference architecture that is elaborated upon throughout the rest of the book. Chapter 2 delves into the evaluation of CIR systems, focusing on goal-oriented conversational AI systems with a human in the loop. Chapters 3 to 7 present the algorithms and methods for developing the core components of CIR. Chapter 3 focuses on conversational document search, which can be considered a sub-system within CIR. Chapter 4 explores query-focused multi-document summarization algorithms and methods. Chapter 5 explores various neural models for conversational machine comprehension, enabling systems to generate direct answers to user queries based on retrieved query-relevant documents. Chapter 6 details neural approaches to conversational question answering over knowledge bases, a fundamental component of CIR systems. Chapter 7 explores techniques and models aimed at equipping CIR systems with the capability to proactively lead human-machine conversations. Chapter 8 provides a review of various commercial systems for CIR and related tasks, offering an overview of the research landscape.

In conclusion, this book serves as a valuable resource for researchers, practitioners, and enthusiasts in the field of CIR. It provides a comprehensive overview of recent advancements in neural approaches, highlighting their transformative impact on NLP and conversational AI. By exploring the development of core CIR modules, evaluation techniques, and proactive conversation management, this book offers insights into the future of information retrieval, paving the way for more human-centric and intuitive interactions with technology.

Weight: 506g
Dimension: 235 x 155 (mm)
ISBN-13: 9783031230790
Edition number: 1st ed. 2023