Call for Papers
We invite submissions describing innovations and implementations that focus on understanding how to optimize human-computer cooperation and minimize human effort along the data science pipeline in a wide range of data science tasks and real-world applications.
Topics of interest to this workshop include, but are not limited to:
Supervised, self-supervised, transfer, and unsupervised learning with human-in-the-loop
Human intervention and consequences on model bias, overfitting, and fairness
Human and computer cooperation in data cleaning, preparation, and representation for NLP tasks
Optimizing human effort in data labeling
Crowdsourcing in NLP
Human-in-the-loop during model evaluation and model improvement
Multi-modal human-in-the-loop issues in NLP in conjunction with data from other modalities such as social media, knowledge graphs, speech, image, video
Enabling non-experts to build advanced NLP models
Formal and higher abstractions for human-machine interaction in NLP tasks
User interfaces for data preprocessing, labeling, and NLP model building, evaluation, and interpretation
Human-in-the-loop issues in implementing, using, evaluating, and deploying NLP models in specific applications (e.g., medical, scientific, legal, business, digital humanities, creative).
Authors are invited to submit either of the following:
Original, unpublished research papers that are not being considered for publication in any other forum. Research papers are limited to six pages in length, excluding references.
Short papers of late-breaking work and work in progress. Abstracts are limited to two pages.
Authors of both types of papers will need to present these papers at DaSH.
Submissions to the workshop must be in PDF and should follow the NAACL paper formatting. All submissions should be anonymized to facilitate double-blind reviewing. To submit a paper, please access the submission link (coming soon).
One Best Paper Award and One Best Student Paper Award will be selected!
The presentation format and schedule will be announced before the camera-ready deadline.