Workshop on Towards Semi-Supervised and Reinforced Task-Oriented Dialog Systems
Co-located with EMNLP 2022
December 7, 2022
Overview
Building task-oriented dialog (TOD) systems remains a label-intensive, time-consuming task for the reliance on manually labeled dialog data and annotated task-related knowledge base. However, unlabeled data are often easily available in many forms such as human-to-human dialogs, open-domain text corpus, and unstructured knowledge documents. Moreover, the nature of sequential decision making is not fully considered in current TOD systems.
The purpose of this workshop is to invite researchers from both academia and industry to share their perspectives on building semi-supervised and reinforced TOD systems, discuss challenges and advance the field in joint effort. A shared task and challenge will be organized for benchmarking and stimulating relevant researches, with a newly released large-scale, multi-domain TOD dataset which consists of 100K real-world dialogs. The Challenge consists of two tracks:
Topics of interest
We welcome submissions from all levels of methodologies, algorithms, models, system developments, applications and datasets towards semi-supervised and reinforced TOD systems. All types of semi-supervised techniques are welcome, such as, to name a few, pre-training, self-training, self-supervised, weakly-supervised, transfer learning for zero-shot or few-shots, latent-variable modeling, and domain adaptation, and data augmentation. Both online and offline reinforcement learning techniques are welcome. Topics of interest include, but are not limited to, the following:
Important Dates
September 17, 2022: Regular paper due (via softconf)
October 2, 2022: ARR commitment due (via OpenReview)
October 9, 2022: Notification of paper acceptance
October 16, 2022: Camera-ready papers due
December 7, 2022: EMNLP 2022 Workshop (hybrid)
Submissions
We accept long papers (8 pages plus unlimited references), short papers (4 pages plus unlimited references), and system description papers for the challenge (in either long or short formats), which should conform to ACL submission information.
Authors can submit a paper up to 2-8 pages long, with unlimited reference and supplementary material pages. The accepted papers can opt to be non-archival, i.e., the work could be published elsewhere before or after the workshop.
All submissions will be peer-reviewed by at least two reviewers. The reviewing process will be double-blinded at the level of the reviewers. Authors are responsible for anonymizing the submissions.
SereTOD Workshop will accept regular submissions through the softconf system, as well as commitment of already reviewed papers through the ACL Rolling Review (ARR) system.
July 15, 2022: ARR paper due (submit via ARR)
September 17, 2022: Regular paper due (submit via softconf)
October 2, 2022: ARR commitment due (commit via OpenReview)
Organizers
Zhijian Ou, Tsinghua University
Junlan Feng, China Mobile
Juanzi Li, Tsinghua University
For any questions, please feel free to contact: seretod2022 (at) gmail (dot) com
Looking forward to your participation!