This repository contains the data, scripts and baseline codes for DSTC9 Track 1.

This challenge track aims to support frictionless task-oriented conversations, where the dialogue flow does not break when users have requests that are out of the scope of APIs/DB but potentially are already available in external knowledge sources. Track participants will develop dialogue systems to understand relevant domain knowledge, and generate system responses with the relevant selected knowledge.

Organizers: Seokhwan Kim, Mihail Eric, Behnam Hedayatnia, Karthik Gopalakrishnan, Yang Liu, Chao-Wei Huang, Dilek Hakkani-tur

• Task Description Paper
• Track Proposal
• Challenge Registration
• Data Formats
• Baseline Details
• Frequently Asked Questions

If you want to publish experimental results with this dataset or use the baseline models, please cite the following article:

@article{kim2020domain,
  title={Beyond Domain APIs: Task-oriented Conversational Modeling with Unstructured Knowledge Access},
  author={Seokhwan Kim and Mihail Eric and Karthik Gopalakrishnan and Behnam Hedayatnia and Yang Liu and Dilek Hakkani-Tur},
  journal={arXiv preprint arXiv:2006.03533}
  year={2020}
}

NOTE: This paper reports the results with an earlier version of the dataset and the baseline models, which will differ from the baseline performances on the official challenge resources.

This challenge track decouples between turns that could be handled by the existing task-oriented conversational models with no extra knowledge and turns that require external knowledge resources to be answered by the dialogue system. We focus on the turns that require knowledge access as the evaluation target in this track by the following three tasks:

Participants will develop systems to generate the outputs for each task. They can leverage the annotations and the ground-truth responses available in the training and validation datasets.

In the test phase, participants will be given a set of unlabeled test instances. And they will submit up to 5 system outputs for all three tasks.

NOTE: For someone who are interested in only one or two of the tasks, we recommend to use our baseline system for the remaining tasks to complete the system outputs.

Each submission will be evaluated in the following task-specific automated metrics first:

To consider the dependencies between the tasks, the scores for knowledge selection and knowledge-grounded response generation are weighted by knowledge-seeking turn detection performances. Please find more details from scores.py.

The final ranking will be based on human evaluation results only for selected systems according to automated evaluation scores. It will address the following aspects: grammatical/semantical correctness, naturalness, appropriateness, informativeness and relevance to given knowledge.

In this challenge track, participants will use an augmented version of MultiWoz 2.1 which includes newly introduced knowledge-seeking turns. All the ground-truth annotations for Knowledge-seeking Turn Detection and Knowledge Selection tasks as well as the agent's responses for Knowledge-grounded Response Generation task are available to develop the components on the training and validation sets. In addition, relevant knowledge snippets for each domain or entity are also provided in knowledge.json.

In the test phase, participants will be evaluated on the results generated by their models for two data sets: one is the unlabeled test set of the augmented MultiWoz 2.1, and the other is a new set of unseen conversations which are collected from scratch also including turns that require knowledge access. To evaluate the generalizability and the portability of each model, the unseen test set will be collected on different domains, entities and locales than MultiWoz.

Data and system output format details can be found from data/README.md.

• Training data released: Jun 15, 2020
• Test data released: Sep 21, 2020
• Entry submission deadline: Sep 28, 2020
• Objective evaluation completed: Oct 12, 2020
• Human evaluation completed: Oct 19, 2020

• Participation is welcome from any team (academic, corporate, non profit, government).
• The identity of participants will NOT be published or made public. In written results, teams will be identified as team IDs (e.g. team1, team2, etc). The organizers will verbally indicate the identities of all teams at the workshop chosen for communicating results.
• Participants may identify their own team label (e.g. team5), in publications or presentations, if they desire, but may not identify the identities of other teams.
• Participants are allowed to use any external datasets, resources or pre-trained models.

• To join the mailing list: visit https://groups.google.com/a/dstc.community/forum/#!forum/list/join

• To post a message: send your message to [email protected]

• To leave the mailing list: visit https://groups.google.com/a/dstc.community/forum/#!forum/list/unsubscribe

Please feel free to contact: seokhwk (at) amazon (dot) com