AAAI 2020 Workshop on Interactive and Conversational Recommendation Systems (WICRS)

Overview

Recent years have seen an increase in work in interactive, sequential (e.g., session-based) interactions with recommender systems. Furthermore, the rise of conversational AI-based assistants in the form of Apple’s Siri, Amazon’s Alexa, and the Google Assistant have invigorated interest in dialog-based sequential interaction, often with a limited degree of personalization.

Workshop Format

Full day. The format will consist of four sessions, all sessions start with an invited talk. The first three sessions will follow with three or four twelve minute presentations of accepted workshop papers. The fourth session will conclude the workshop with a panel discussion drawing panelists from academia and industry focusing on the frontiers of interactive and conversational recommendation. Contributed papers not selected for oral presentation will be presented in three minute spotlight talks. All papers are invited to be presented as posters during morning and afternoon coffee breaks.

Registration and Attendance

Update Feb. 3, 2020: it appears the workshop attendance is full, please contact Stephanie Le (le@aaai.org) about registering for on-site registrations.

While the early registration deadline (December 13) has passed, late registration is still open (deadline is January 10), and on-site registration is available after January 10. No invitation required.

See https://aaai.org/Conferences/AAAI-20/registration/ for registration instructions and https://aaai.org/Conferences/AAAI-20/hotel-and-travel/ for hotel and travel suggestions.

Dates and Location

Paper submission (extended): Nov. 18, 2019, 11:59pm AOE -- Anywhere On Earth.
Notifications of acceptance: Dec. 4, 2019
Camera-ready due: Dec. 21, 2019
Workshop date: Feb. 8, 2020 (full day)
Room and Location:

Clinton, 2nd floor of hotel

Hilton New York Midtown, New York, USA

Contact: wicrs20@easychair.org

Schedule

Workshop Room: Clinton, 2nd floor

Session 1

8:50 - 9:00 Opening Remarks

9:00 - 9:40 Invited Talk - Michelle Zhou, “You Really Get Me”: Conversational AI That Can Truly Understand and Help Users [slides] [demo]

9:40 - 9:52 Contributed Oral - Improving recommendation by deep latent factor-based explanation, Sixun Ouyang and Aonghus Lawlor

9:52 - 10:04 Contributed Oral - Explanations within Conversational Recommendation Systems: Improving Coverage through Knowledge Graph Embedding [slides], Gustavo Polleti, Hugo Neri and Fabio Cozman

10:04 - 10:16 Contributed Oral - Deep Critiquing for VAE-based Recommender Systems, Kai Luo, Hojin Yang, Ga Wu and Scott Sanner

10:16 - 10:30 Spotlight talks (3 min each)

MSET: Multi-Scale Exact Test for Detecting Distinctive Periods in Choice Behavior, Kei Shimonishi and Hiroaki Kawashima
What Do You Mean I’m Funny? Personalizing the Joke Skill of a Voice-Controlled Virtual Assistant, Alejandro Mottini and Amber Roy Chowdhury
[Update 2/6: CANCELLED due to Coronavirus outbreak] Why Do Neural Response Generation Models Prefer Universal Replies?, Bowen Wu, Nan Jiang, Zhifeng Gao, Mengyuan Li, Zongsheng Wang, Suke Li, Wenge Rong and Baoxun Wang

10:30 - 11:00 Coffee break & poster session - all papers

Session 2

11:00 - 11:40 Invited Talk - Maarten de Rijke

11:40 - 11:52 Contributed Oral - Planning for Goal-Oriented Dialogue Systems, Christian Muise, Tathagata Chakraborti, Shubham Agarwal, Ondrej Bajgar, Arunima Chaudhary, Luis Lastras, Josef Ondrej, Miroslav Vodol and Charlie Wiecha

11:52 - 12:04 Contributed Oral - Answering Comparative Questions with Web-based Arguments, [slides] Alexander Bondarenko, Matthias Hagen, Meriem Beloucif, Chris Biemann and Alexander Panchenko

12:04 - 12:16 Contributed Oral - Multi-Gradient Descent for Multi-Objective Recommender Systems, Nikola Milojkovic, Diego Antognini, Giancarlo Bergamin and Claudiu Musat

12:16 - 1:45 Lunch

Session 3

1:45 - 2:25 Invited Talk - Craig Boutilier, Interactive Recommenders and Rich User Modeling

2:25 - 2:37 Contributed Oral - A Bayesian Approach to Conversational Recommendation Systems [slides], Francesca Mangili, Denis Broggini, Alessandro Antonucci, Marco Alberti and Lorenzo Cimasoni

2:37 - 2:49 Contributed Oral - Active Learning in Conversational Recommendation Systems with Multi-level User Preferences, Yuheng Bu and Kevin Small

2:49 - 3:01 Contributed Oral - Recommendation by Joining a Human Conversation, Boris Galitsky and Dmitry Ilvovsky

3:01 - 3:13 Contributed Oral - Multimodal Conversational Recommendation Systems, Noriaki Kawamae

3:13 - 3:30 Spotlight talks (3 min each)

Make Your Customers Happy Again: AI and NLP for a customer complaint management platform, Maximilien Kintz, Claudia Dukino, Matthias Blohm and Marc Hanussek
Question Generation by Transformers, Kettip Kriangchaivech and Artit Wangperawong
An Attention-Based Speaker Naming Method for Online Adaptation in Non-Fixed Scenarios, Jungwoo Pyo, Joohyun Lee, Youngjune Park, Tien-Cuong Bui and Sang Kyun Cha
Recurrent Point Processes for Dynamic Review Models, Kostadin Cvejoski, Ramses J. Sanchez, Bogdan Georgiev, Jannis Schuecker, Christian Bauckhage and Cesar Ojeda

3:30 - 4:00 Coffee break & poster session - all papers

Session 4

4:00 - 4:40 Invited Talk - Zhou Yu, Seamless Natural Communication between Humans and Machines

4:40 - 5:30 Panel Discussion - Michelle Zhou, Maarten de Rijke, Craig Boutilier, Zhou Yu. Moderator: Scott Sanner

5:30 - 5:40 Closing Remarks

Invited Speaker - Michelle Zhou

Michelle Zhou, Juji Inc.

Title: “You Really Get Me”: Conversational AI That Can Truly Understand and Help Users

Abstract: Have you watched the movie Her? Have you ever wished to have an AI companion like Samantha, who could tell you what you really are, whom your best teammate may be, and which career path would be best for you? In this talk, Michelle will present a framework for building hyper-personalized, conversational Artificial Intelligent (AI) agents who can deeply understand users and responsibly guide user behavior in both virtual and real world. Through live demos, she will highlight two technical advances of the framework: (1) evidence-based personality inference and (2) model-based conversation generation. Michelle will describe real-world applications of these agents and discuss the wider implications of enabling hyper-personalized conversational AI for businesses and individuals.

Bio: Dr. Michelle Zhou is a Co-Founder and CEO of Juji, Inc., a high-tech startup located in Silicon Valley, specializing in building responsible and empathetic Artificial Intelligence agents that can deeply understand users and guide their behavior based on their psychological characteristics. Prior to starting Juji, Michelle led the User Systems and Experience Research (USER) group at IBM Research – Almaden and then the IBM Watson Group. Michelle’s expertise is in the interdisciplinary area of intelligent user interaction (IUI), including conversational systems and personality analytics. She has published over 100 peer-reviewed, refereed articles and over 45 patents. Michelle is currently the Editor-in-Chief of ACM Transactions on Interactive Intelligent Systems (TiiS) and an Associate Editor of ACM Transactions on Intelligent Systems and Technology (TIST). She received a Ph.D. in Computer Science from Columbia University and is an ACM Distinguished Scientist.

Invited Speaker - Zhou Yu

Zhou Yu, UC Davis

Title: Seamless Natural Communication between Humans and Machines

Abstract: Dialog systems such as Alexa and Siri are everywhere in our lives. They can complete tasks such as booking flights, making restaurant reservations and training people for interviews. However, currently deployed dialog systems are rule-based and cannot generalize to different domains, let alone flexible dialog context tracking. We will first discuss how to design studies to collect realistic dialogs through a crowdsourcing platform. Then we introduce a dialog model that utilizes limited data to achieve good performance by leveraging multi-task learning and semantic scaffolds. We further improve the model's coherence by tracking both semantic actions and conversational strategies from dialog history using finite-state transducers. Finally, we analyze some ethical concerns and human factors in dialog system deployment. All our work comes together to build seamless natural communication between humans and machines.

Bio: Zhou Yu is an Assistant Professor at the UC Davis Computer Science Department. She obtained her Ph.D. from Carnegie Mellon University in 2017. Dr. Yu has built various dialog systems with major practical impacts, such as a job interview training system, a depression screening system, an Alexa social chatbot and a second language learning system. Her research interest includes dialog systems, language understanding and generation, vision and language, human-computer interaction and social robots. Dr. Yu's work earned a 2019 ACL best paper award nomination. She was recognized in the Forbes 2018 30 under 30 in Science, and won the 2018 Amazon Alexa Prize

Invited Speaker - Maarten de Rijke

Maarten de Rijke, University of Amsterdam

Title: TBA

Bio: Maarten de Rijke is University Professor of Artificial Intelligence and Information Retrieval at the University of Amsterdam and Director of the Information and Language Processing Systems lab. De Rijke’s research strives to build ever more intelligent technology to connect people to information. His team pushes the frontiers of search engines, recommender systems and conversational assistants. They also investigate the influence of the technology they develop on society. De Rijke is the director of the national Innovation Center for Artificial Intelligence.

Invited Speaker - Craig Boutilier

Craig Boutilier, Google Research

Title: Interactive Recommenders and Rich User Modeling

Abstract: While recommender systems have had a significant impact on the way we discover and consume content, products and services, there remains a significant gap in the ability of recommenders to truly act in a user's best interests. This is, in part, due to the impoverished nature of user models used by current recommender technology. In this talk, I will outline a few dimensions in which user models should be expanded in order for recommenders to better serve a user's interest—including rich preference and context representations. I will also discuss how rich interactive (including conversational) models, by increasing the bandwidth of the communication channel with users, can facilitate the construction of such models. Along the way, I will highlight some of the interesting research challenges that emerge in this endeavor.

Bio: Dr. Craig Boutilier, Principal Research Scientist at Google, has made seminal contributions to research on decision-making under uncertainty, game theory, and computational social choice. He is a pioneer in applying decision-theoretic concepts in novel ways in a variety of domains including (single- and multi-agent) planning and reinforcement learning, preference elicitation, voting, matching, facility location, and recommender systems. His recent research continues to significantly influence the field of computational social choice through the novel computational and methodological tools he introduced and his focus on modeling realistic preferences. In addition to his reputation for outstanding research, Dr. Boutilier is also recognized as an exceptional teacher and mentor.

Call for Papers

This workshop is a forum to present and discuss novel research directions in interactive and conversational recommender systems as well as the constituent AI technologies that represent the next generation of recommender systems and personalized, conversational assistants. The workshop aspires to bring together AI researchers from recommender systems, machine and reinforcement learning, dialog systems, natural language processing, human computer interaction, psychology and econometrics for a day of research presentations and open discussion about the future of this high impact and highly cross-disciplinary research area.

Submission Procedure

We welcome previously unsubmitted work, papers submitted to the main AAAI conference, and papers reporting research already published provided they align well with the workshop topic.

Three types of submissions are solicited:

full-length papers (up to 7 pages + 1 page for references in AAAI format)
challenge or position papers (2 pages + 1 page for references in AAAI format)
already published papers (1 page: an abstract in AAAI format with a link to the full paper)

Paper Submissions should be made through the workshop EasyChair web site:

https://easychair.org/conferences/?conf=wicrs20