TUESDAY, JUNE 6 – ARRIVAL

18:00 – 19:00 Registration

19:00 – 21:00 Welcome reception

Activity to Combat Jet Lag – Sunset Watching

Go to the upper deck above the entrance and take a seat!

 

WEDNESDAY, JUNE 7 – DAY 1 (full day at Summit Inn)

7:30 – 8:50 Breakfast on your own

9:00 -9:15 Welcome

9:15– 10:15 Keynote 1 - Dilek Hakkani-Tur, Google


Recent Advances and Challenges on Deep Learning for Goal Oriented Conversational Understanding


Recent advances in deep learning based approaches enabled exciting new research frontiers for goal-oriented conversation systems. In this talk, I will present an end-to-end dialog system, with components for language understanding, dialogue state tracking, policy, and language generation. These can be independently built and jointly optimized for dialogue quality and efficient task completion using supervised, unsupervised, or reinforcement learning. I will focus on novel aspects of each component, and highlight remaining issues and challenges towards building human-level conversational systems.

 

10:15 -10:30 SOFTBANK ROBOTICS coffee break

10:30 – 12:00 Paper Session 1 Chair: Diane Litman

10:30 – 10:50 Nurul Lubis, Sakriani Sakti, Koichiro Yoshino and Satoshi Nakamura, Eliciting positive emotional impact in dialogue response selection


10:50 – 11:10 Zahra Rahimi, Diane Litman and Susannah Paletz, Acoustic-prosodic entrainment in multi-party spoken dialogues: Does simple averaging extend existing pair measures properly?

11:10 – 11:30 Svetlana Stoyanchev, Soumi Maiti and Srinivas Bangalore, Predicting interaction quality in customer service dialogs

11:30 – 11:50 Pierrick Milhorat, Divesh Lala, Koji Inoue, Zhao Tianyu, Masanari Ishida, Katsuya Takanashi, Shizuka Nakamura and Tatsuya Kawahara, A conversational dialogue manager for a humanoid robot

12:00 – 13:00 Lunch

13:00 – 16:55 WOCHAT Special Session

13:00 Special Session Introduction – David Traum

13:20 – 15:00 Oral WOCHAT session – Chair: Wolfgang Minker

13:20 Koh Mitsuda, Ryuichiro Higashinaka and Yoshihiro Matsuo, What information should a dialog system understand? Collection and analysis of perceived information in chat-oriented dialog Long Paper 

13:40 Emer Gilmartin, Carl Vogel, Nick Campbell, Benjamin Cowan, Chunks in multiparty conversation - building blocks for extended social talk Short paper

13:55 Jacqueline Brixey and David Novick, Building rapport with extraverted and introverted agents Long Paper 


14:15 Geetanjali Rakshit, Kevin Bowden, Lena Reed, Amita Misra and Marilyn Walker, Debbie the Debate Tron of the future Short Paper  


14:30 Hiroaki Sugiyama, Toyomi Meguro and Ryuichiro Higashinaka. Automatic evaluation of chat-oriented dialogue systems using large-scale multi-references Long paper


14:50 – 15:00 Wrapup and concluding remarks from the oral session


15:00 - 15:15 INTERACTIONS coffee break


15:15 – 17:00 Chat session continued  

15:15 Shared task report and continuation plans including JSALT proposal – Zhou Yu

15:35 Panel – Moderator, Joseph Mariani

Kevin Bowden, Laurence Devillers, Satoshi Nakamura, Alex Rudnicky, Marilyn Walker, Zhou Yu


16:35 – 17:00  Plenary final WOCHAT discussion including plans for future WOCHAT efforts  – Ron Artstein


17:00 – 18:00 Tree House visits

18:00 – 23:00 Dinner



THURSDAY, JUNE 8 – DAY 2 (partial day at Summit Inn & event at Fallingwater/winery)

7:30 – 8:50 Breakfast on your own

9:00 - 10:00 Keynote 2 Joanna Bryson, Princeton and U. Bath

Can language use by artefacts be ethical?

Natural language is expected to be the dominant interface to AI in the near future, in fact this time may already be upon us. While more accessible to most populations than conventional keyboard and screen-based interfaces, natural language carries with it a number of ethical challenges.

  • privacy and security: as we open our human communicative contexts to machines that perceive what we do, we allowing intimate information to be communicated over the Internet.

  • transparency: using a more human-seeming interface can create a greater illusion of human-like capacities, which may be a moral hazard.

  • fairness: learning natural language requires absorbing human culture, which can include stereotypes we wish weren't there.

I will go over each of these, devoting the most time to recent results on bias in NLP in work with Aylin Caliskan and Arvind Narayanan (Science, 2017).  I will also discuss measures that can be taken to ameliorate each of these problems, many surprisingly simple, but all requiring good faith effort, potentially backed with government regulation enforcing industry-agreed standards.


10:00 10:15 FACEBOOK coffee break

10:30 – 12:00 Multi-domain Special Session Dilek Hakkani-Tur, Maxine Eskenazi

10:30 – 10:50 Kyusong Lee, Tiancheng Zhao, Stefan Ultes, Lina Rojas-Barahona, Eli Pincus, David Traum and Maxine Eskenazi, An assessment framework for DialPort

10:50 – 11:10 Alexandros Papangelis, Stefan Ultes and Yannis Stylianou, Domain complexity and policy learning in task-oriented dialogue systems

11:10 – 11:30 Alexandros Papangelis and Yannis Stylianou, Single-model multi-domain dialogue management with deep learning

11:30 – 11:50 Michael Wessel, Girish Acharya, James Carpenter and Min Yin. An Ontology-Based Dialogue Management System for Virtual Personal Assistants

12:00 – 13:00 Lunch

13:45 Depart for Fallingwater

14:30 Fallingwater check-in

15:00 Fallingwater tours begin

17:00 Depart from Fallingwater for Christian Klay Winery

17:30 – 21:30 Hay ride and banquet at Christian Klay Winery

21:30 Depart Christian Klay Winery to return to Summit Inn (+ bonfire and s’mores)



FRIDAY, JUNE 9 – DAY 3 (half day at Summit Inn, departure)

7:30 – 8:50 Breakfast on your own

9:00 - 10:15 Robots, Evaluation and Ethics Special Session Laurence Devillers, Amit Kumar Pandey

9:00 – 9:20 Juliana Miehle, Ilker Bagci, Wolkgang Minker and Stefan Ultes, A social companion and conversation partner for the elderly

9:20 – 9:40 Melanie Garcia, Guillaume Dubuisson-Duplessis, Gabrielle Pittaro, Lucile Béchade and Laurence Devillers. Towards metrics of Evaluation of Pepper robot as a Social Companion for Elderly People

9:40 – 10:00 Pascale Fung, Farhad Bin Siddique, Onno Kampmann and Yang Yang, User personality adapting conversational agent

10:00 – 10:20 Robots, Evaluation and Ethics Panel Discussion Laurence Devillers, Joana Bryson

10:20 - 10:35 AMAZON ALEXA coffee break

10:35 – 12:10 Poster session Chair: Wolfgang Minker

Louisa Pragst, Wolfgang Minker and Stefan Ultes, Exploring the applicability of verbosity and indirectness in dialogue management

Yukitoshi Murase, Koichiro Yoshino, Masahiro Mizukami and Satoshi Nakamura, Feature inference based on label propagation on Wikidata graph for DST

Bethany Lycan and Ron Artstein, Direct and mediated interaction with a Holocaust survivor

Tetsuro Takahashi and Hikaru Yokono, Two person dialogue corpus made by multiple crowd-workers

Patrick Ehrenbrink and Stefan Hillmann, Comparing priming effects of visual and textual task representations - texts can influence users' utterances

Stefan Ultes, Juliana Miehle and Wolfgang Minker, On the applicability of a user satisfaction-based reward for dialogue policy learning

Manex Serras, María Inés Torres and Arantza Del Pozo, Attention-based neural user model for goal oriented spoken dialogue systems

Robin Ruede, Markus Müller, Sebastian Stuker and Alex Waibel, Yeah, right, uh-huh: A deep learning backchannel predictor

Zhou Yu, Vikram Ramanarayanan, Patrick Lange and David Suendermann-Oeft, An open-source dialog system with real-time user engagement coordination for job interview training applications

12:10 -12:30 Conclusion and Presentation of IWSDS 2018

12:30 Buffet lunch

13:30 Bus departs for Pittsburgh airport