Jianfeng Gao

[introductory/intermediate] Neural Approaches to Conversational Information Retrieval

Summary

A conversational information retrieval (CIR) system is an information retrieval (IR) system with a conversational interface which allows users to interact with the system to seek information via multi-turn conversations of natural language (in spoken or written form). This tutorial surveys recent advances in CIR, focusing on neural approaches that have been developed in the last few years. We present (1) a typical architecture of a CIR system, (2) new tasks and applications which arise from the needs of developing such a system, in comparison with traditional keyword-based IR systems, (3) new methods of conversational question answering, and (4) case studies of several CIR systems developed in research communities and industry.

Syllabus

This tutorial will include four-hour materials scheduled as in the following six parts:

Part 1: Introduction to Conversational Information Retrieval (CIR)

This part motivates the research of CIR by reviewing the studies on how people search and how information seeking can be cast as human-machine conversations.

Part 2: Conversational Search Systems

This part discusses the core components of CIR systems, focusing on the new challenges raised from the conversational nature in comparison with classic, keyword-based, IR systems. There are three sections in this part. The first section describes the general framework of a CIR system; the remaining two present two core CIR components, conversational query understanding and conversational document ranking, respectively.

Part 3: Interactive Information Retrieval Methods

This part presents a set of methods that aim to make CIR systems play a more active role in human-machine interactions of information seeking. These methods collectively allow a CIR system to actively help a user clarify her search intent by asking clarification questions, guide the user to discover and learn new knowledge by suggesting and recommending related information and search in new directions, and synthesize retrieved content to form a new understanding. Finally, we present the challenges and solutions when evaluating such an interactive CIR system.

Part 4: Conversational QA over Texts

This part presents recent developments in conversational Question Answering (QA) systems that allow users to query a document collection in natural language. Such text-QA agents are much easier to use in mobile devices than traditional search engines in that they provide concise, direct answers to user queries, as opposed to a list of document links.

Part 5: Conversational QA over Structured Databases

This part presents recent developments in conversational Question Answering over structured databases, i.e. Knowledge Bases (KBs).

Part 6: Case Studies of Commercial Systems

The last part of this tutorial reviews a variety of commercial systems for CIR and related tasks. Due to the proprietary nature of many of these systems, we limit our review to summarizing published material about the systems.

References

• [1] Ricardo Baeza-Yates, Berthier Ribeiro-Neto, et al. 1999. Modern information retrieval. Vol. 463. ACM Press, New York.
• [2] Marcia J Bates. 1979. Information search tactics. Journal of the American Society for information Science 30, 4 (1979), 205–214.
• [3] Marcia J Bates. 1989. The design of browsing and berrypicking techniques for the online search interface. Online review 13, 5 (1989), 407–424.
• [4] Helen M Brooks, Penny J Daniels, and Nicholas J Belkin. 1986. Research on information interaction and intelligent information provision mechanisms. Journal of Information Science 12, 1-2 (1986), 37–44.
• [5] Philipp Christmann, Rishiraj Saha Roy, Abdalghani Abujabal, Jyotsna Singh, and Gerhard Weikum. 2019. Look before you Hop: Conversational Question Answering over Knowledge Graphs Using Judicious Context Expansion. In Proceedings of the 28th ACM International Conference on Information and Knowledge Management. 729–738.
• [6] Kevyn Collins-Thompson, Preben Hansen, and Claudia Hauff. 2017. Search as learning (dagstuhl seminar 17092). In Dagstuhl reports, Vol. 7. Schloss DagstuhlLeibniz-Zentrum fuer Informatik.
• [7] W. Bruce Croft. 2019. The Importance of Interaction for Information Retrieval. In Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. ACM, New York, NY, USA, 1–2.
• [8] J. Shane Culpepper, Fernando Diaz, and Mark Smucker. 2018. Research Frontiers in Information Retrieval: Report from the Third Strategic Workshop on Information Retrieval in Lorne (SWIRL 2018). SIGIR Forum 52, 1 (2018), 34–90.
• [9] Douglass R Cutting, David R Karger, Jan O Pedersen, and John W Tukey. 2017. Scatter/gather: A cluster-based approach to browsing large document collections. In ACM SIGIR Forum, Vol. 51. ACM, New York, NY, USA, 148–159.
• [10] Rand Fishkin. 2019. Less than Half of Google Searches Now Result in a Click. sparktoro.com/blog/less-than-half-of-google-searches-now-result-in-a-click/. Accessed:2020-01-23.
• [11] Jianfeng Gao, Michel Galley, and Lihong Li. 2018. Neural Approaches to Conversational AI. In The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval. 1371–1374.
• [12] Jianfeng Gao, Michel Galley, Lihong Li, et al. 2019. Neural approaches to conversational AI. Foundations and Trends® in Information Retrieval 13, 2-3, 127–298.
• [13] Mohit Iyyer, Wen-tau Yih, and Ming-Wei Chang. 2017. Search-based neural structured learning for sequential question answering. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 1821–1831.
• [14] Ying Ju, Fubang Zhao, Shijie Chen, Bowen Zheng, Xuefeng Yang, and Yunfeng Liu. 2019. Technical report on Conversational Question Answering. arXiv preprint arXiv:1909.10772 (2019).
• [15] Michael E Lesk and Gerard Salton. 1969. Interactive search and retrieval methods using automatic information displays. In Proceedings of the May 14-16, 1969, spring joint computer conference. ACM, 435–446.
• [16] Thomas Müller, Francesco Piccinno, Massimo Nicosia, Peter Shaw, and Yasemin Altun. 2019. Answering Conversational Questions on Structured Data without Logical Forms. arXiv preprint arXiv:1908.11787 (2019).
• [17] Robert N Oddy. 1977. Information retrieval through man-machine dialogue. Journal of documentation 33, 1 (1977), 1–14.
• [18] Filip Radlinski and Nick Craswell. 2017. A theoretical framework for conversational search. In Proceedings of the 2017 conference on conference human information interaction and retrieval. ACM, 117–126.
• [19] Ben Shneiderman, Don Byrd, and W Bruce Croft. 1997. Clarifying search: A user-interface framework for text searches. D-lib magazine 3, 1 (1997), 18–20.
• [20] Pei-Hao Su, Nikola Mrkšić, Iñigo Casanueva, and Ivan Vulić. 2018. Deep learning for conversational AI. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Tutorial Abstracts. 27–32.
• [21] Anastasios Tombros, Mark Sanderson, and Phil Gray. 1998. Advantages of query biased summaries in information retrieval. In SIGIR, Vol. 98. 2–10.
• [22] Tsung-Hsien Wen, Pei-Hao Su, Paweł Budzianowski, Iñigo Casanueva, and Ivan Vulić. 2019. Data Collection and End-to-End Learning for Conversational AI. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP): Tutorial Abstracts. ACL.
• [23] Wei Wu and Rui Yan. 2019. Deep Chit-Chat: Deep Learning for Chatbots. In Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. 1413–1414.
• [24] Li Zhou, Jianfeng Gao, Di Li, and Heung-Yeung Shum. 2018. The design and implementation of XiaoIce, an empathetic social chatbot. arXiv preprint arXiv:1812.08989 (2018).

Pre-requisites

Ricardo Baeza-Yates, Berthier Ribeiro-Neto, et al. 1999. Modern information retrieval. Vol. 463. ACM Press, New York.

Short bio

Jianfeng Gao is a Distinguished Scientist and Vice President of Microsoft. He is the head of the Deep Learning group at Microsoft Research, leading the development of AI systems for natural language processing, Web search, vision language understanding, dialogue, and business applications. He is an affiliate professor of Computer Science & Engineering at University of Washington, an IEEE Fellow, and a Distinguished Member of ACM.