Conversational AI: Today's Practice and Tomorrow's Potential
About the workshop
In the span of only a few years, conversational systems have become commonplace. Every day, millions of people use natural-language interfaces such as Siri, Google Now, Cortana, Alexa, Facebook M and others via in-home devices, phones, or messaging channels such as Messenger, Slack, Skype, among others. At the same time, interest among the research community in conversational systems has blossomed; For supervised and reinforcement learning, conversational systems often serve as both a benchmark task and an inspiration for new ML methods at conferences which don't focus on speech and language per se, such as NIPS, ICML, IJCAI, and others. Research community challenge tasks are proliferating, including the sixth Dialog Systems Technology Challenge (DSTC6), the Amazon Alexa prize, and the Conversational Intelligence Challenge live competition at NIPS 2017. Now more than ever, it is crucial to promote cross-pollination of ideas between academic research centers and industry. The goal of this workshop is to bring together researchers and practitioners in this area, to clarify impactful research problems, to share findings from large-scale real-world deployments, and to generate new ideas for future lines of research. This workshop will include invited talks from academia and industry, contributed work, and open discussion. In these talks, senior technical leaders from many of the most popular conversational services will give insights into real usage and challenges at scale. An open call for papers will be issued, and we will prioritize forward-looking papers that propose interesting and impactful contributions. We will end the day with an open discussion, including a panel consisting of academic and industrial researchers.
Speakers
Schedule
- 9:00-9:10Opening Remarks - Alborz Geramifard (Amazon)
- 9:10-9:50Invited Talk: - Satinder Singh (UMich)
- 9:50-10:10Accepted Talk: An Ensemble Model with Ranking for Social Dialogue
- 10:10-10:30Accepted Talk: A Deep Reinforcement Learning Chatbot
- 10:30-11:00Coffee Break / Hangup Posters
- 11:00-11:40Invited Talk: Social and Ethical Aspects of Dialogue Systems - Joelle Pineau (McGill/Facebook)
- 11:40-12:00Accepted Talk: Multi-Domain Adversarial Learning for Slot Filling in Spoken Language Understanding
- 12:00-1:30Lunch Break
- 1:30-2:10Invited Talk: Conversational AI at Scale: The Science behind the Alexa Prize - Ashwin Ram (Amazon)
- 2:10-2:50Invited Talk: Evolution of Siri - Blaise Thomson (Apple)
- 2:50-3:05One Minute Poster Spotlight
- 3:05-3:30Coffee Break / Poster Session
- 3:30-3:50Invited Talk: Neural Models for Text Generation - Asli Celikyilmaz (Microsoft Research)
- 3:50-4:10Invited Talk: Composite task completion bot with Hierarchical RL - Jianfeng Gao (Microsoft Research)Slides ↗
- 4:10-4:30Invited Talk: Smart reply and Implicit semantics - Matt Henderson (Google)Slides ↗
- 4:30-4:50Invited Talk: Conversational Machines: Deep Learning for Goal Oriented Dialogue Systems - Dilek Hakkani-Tur (Google)
- 4:50-5:20Poster Session
- 5:20-6:00Invited Talk: Quantized Dialog - David Nahamoo (IBM)Slides ↗
- 6:00-6:50Panel / Open Discussion - All Organizers (submit questions here and enter your info here for Echo Dot raffle)
- 6:50-7:00Closing Remarks - Jason Williams (Microsoft Research)
Accepted papers
- An Ensemble Model with Ranking for Social DialogueIoannis Papaioannou, Amanda Cercas Curry, Jose L. Part, Igor Shalyminov, Xinnuo Xu, Yanchao, Yu Ondrej Dušek, Verena Rieser, and Oliver Lemon
- A Deep Reinforcement Learning ChatbotIulian V. Serban, Chinnadhurai Sankar, Mathieu Germain, Saizheng Zhang, Zhouhan Lin, Sandeep Subramanian, Taesup Kim, Michael Pieper, Sarath Chandar, Nan Rosemary Ke, Sai Rajeshwar, Alexandre de Brebisson, Jose M. R. Sotelo, Dendi Suhubdy, Vincent Michalski, Alexandre Nguyen, Joelle Pineau, and Yoshua Bengio
- Multi-Domain Adversarial Learning for Slot Filling in Spoken Language UnderstandingBing Liu, and Ian Lane
- Personalization in Goal-oriented DialogChaitanya Joshi, Fei Mi, and Boi Faltings
- Soc2Seq: Social Embedding meets Conversation ModelParminder Bhatia, Marsal Gavaldà, and Arash Einolghozati
- Rasa: Open Source Language Understanding and Dialogue ManagementTom Bocklisch, Joey Faulkner, Nick Pawlowski Rasa, and Alan Nichol
- Just ASK: Building an Architecture for Extensible Self-Service Spoken Language UnderstandingAnjishnu Kumar, Arpit Gupta, Julian Chan, Sam Tucker, Bjorn Hoffmeister, and Markus Dreyer
- Achieving Fluency and Coherency in Task-oriented DialogRashmi Gangadharaiah, Balakrishnan Narayanaswamy, and Charles Elkan
- DialPort: Real-World Data for Academia Spoken Dialog SystemsTiancheng Zhao, Yulun Du, Kyusong Lee, and Maxine Eskenazi
- End-to-End Optimization of Task-Oriented Dialogue Model with Deep Reinforcement LearningBing Liu, Gokhan Tür, Dilek Hakkani-Tür, Pararth Shah, and Larry Heck
- On Evaluating and Comparing Conversational AgentsAnu Venkatesh, Chandra Khatri, Ashwin Ram, Fenfei Guo, Raefer Gabriel, Ashish Nagar, Rohit Prasad, Ming Cheng, Behnam Hedayatnia, Angeliki Metallinou, Rahul Goel, Shaohua Yang, and Anirudh Raju
- Topic-based Evaluation for Conversational BotsFenfei Guo, Angeliki Metallinou, Chandra Khatri, Anirudh Raju, Anu Venkatesh, and Ashwin Ram
- Differentiable lower bound for expected BLEU scoreMaksim Kretov, and Vlad Zhukov
- Learning Robust Dialog Policies in Noisy EnvironmentsMaryam Fazel-Zarandi, Shang-Wen Li, Jin Cao, Jared Casale, David Whitney, and Alborz Geramifard
- Collaboration-based User Simulation for Goal-oriented Dialogue SystemsDevin Didericksen, Oleg Rokhlenko, Kevin Small, and Li Zhou
- End-to-End Offline Goal-Oriented Dialog Policy Learning via Policy GradientLi Zhou, Kevin Small, Oleg Rokhlenko, and Charles Elkan
Call for papers
We invite you to submit your contributions here in the area of conversational AI by November 1st 11:59 PM EST in NIPS style with 2-4 pages with maximum of 8 pages. You do not need to anonymize your submissions. Notifications will be sent out by November 21st . Selected papers will be presented through invited talks or posters (max size of 36 x 48in).
Organizers
