Storyline-only track
Submission of only a 1-page Storyline; non-archival, no proceedings publication, but allows poster presentation for feedback and brainstorming.
Computer Vision/CORE Unranked
ET at ECCV 2026
ECCV 2026 Workshop on Empirical Theory in Representation Learning
The E.T. workshop at ECCV 2026 promotes empirical theory building in deep representation learning, focusing on understanding why deep learning methods work through rigorous empirical evidence rather than pure mathematical theory or state-of-the-art improvements without explanation. It encourages research that proposes, validates, or falsifies hypotheses about neural network behavior, introduces empirical laws, and employs controlled experiments to generalize insights beyond idiosyncratic systems.
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Full paper track
Submission of a 14-page paper (excluding references) in ECCV 2026 format with a 1-page Storyline in the appendix; accepted papers are published in official proceedings and may be selected for oral presentation.
Research areas in scope
01
Research Scope
Propose, validate, and/or falsify hypotheses about the inner workings of deep networksEmpirical observations to inform or inspire theoretical modelsMinimal analytical models that explain observed phenomenaControlled experiments for compiling rigorous empirical evidenceReproduce prior empirical results in simplified or extended settingsIntroduce new experimental tools and methodologies for studying representation learning
Policies worth checking twice
- All submissions must include a 1-page Storyline following the structure at https://jvgemert.github.io/storyline.pdf
- Submissions without a compliant Storyline are at risk of desk rejection
- All submissions are anonymous
- Full papers must follow ECCV 2026 Submission Policies
- Submissions must be made via OpenReview, and authors are required to have an OpenReview profile
- Use of institutional email is preferred for faster approval
- Reviewers assess papers using specific questions focused on empirical rigor, argumentation, and fit to workshop scope
- Storyline components need not all be present — only relevant parts are required
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026