We invite short, non-archival paper submissions focusing on data-centric topics in robot learning, which include but are not limited to the following:
Collection and curation of novel datasets
Investigations on the properties of robot learning datasets
Theoretical or empirical investigations on measures of robot data quality, such as coverage or diversity
Composition and mixing of heterogeneous robotics datasets
Practical strategies for curating data, such as filtering, selection, reweighting/mixing
Tools and frameworks to interpret policy behavior as a function of training data
Active learning and data collection
Self-supervision signals for robotics (pre)-training
Data-centric algorithms towards increased robustness or "model debugging", such as diagnosing biases in training data (e.g., spurious correlations, distribution shifts)
Contributions should make some novel algorithmic, theoretical, benchmarking (e.g., dataset, simulator), or perspective advancement that promotes policy performance and/or our understanding of how behaviors results from training data. Accepted papers and supplementary material will be made available on this workshop website both before the workshop date and subsequently for future visitors unless otherwise requested by authors. Sharing papers in this way does not constitute formal proceedings, i.e., this workshop is a non-archival venue that will not restrict later renditions of the work from being published in archival conferences or journals.
Important Dates:
Submission Portal Opens July 16th, 2025
Submission Deadline August 22nd, 2025
Reviews and Decisions September 1st, 2025
Camera Ready Deadline September 22nd, 2025
Submission Guidelines:
Submissions must be made via OpenReview and formatted according to the CoRL 2025 Conference Template. Papers are recommended to be 6 pages, with a maximum of 8 pages (excluding references, acknowledgments, and optional appendix). Authors are encouraged to include supplementary materials (e.g., videos, code, or datasets) either within a single ZIP file uploaded to OpenReview or via anonymized GitHub links.
The review process will be double blind and non-public. Only accepted papers will be made publicly available after the conclusion of the review process.
All accepted regular papers will have either a short lightning talk or a longer spotlight to advertise their works, followed by a poster session, which will allow more engaged one-on-one conversations between early-career researchers and senior experts. At this time, we are not planning to accommodate any remote presentation options.
Submission Portal: OpenReview