Research papers
Long papers (8 pages) showcasing novel findings, methods, or theoretical advancements.
Computer Vision/CORE Unranked
CVPR 2026
CVPR 2026 Workshop on Bridging Vision, Language, and Action: What’s Missing in Actionable Visual Perception for Robotics
ActiVis@CVPR26 is a workshop focused on bridging computer vision and robotics by developing actionable visual perception systems that enable robots to reason about pose, dynamics, and affordances in real-world environments. It emphasizes co-designed perception-action loops, moving beyond passive vision to task-driven, physically-grounded models that close the loop from pixels to torque.
Key deadlines
Verified September 9, 2026Abstract registration
May 15, 2026
AoE
Full paper
May 15, 2026
AoE
Workshop timeline
Submission and decisions
Abstract registrationKey deadline
May 15, 2026 · AoE
Full paperKey deadline
May 15, 2026 · AoE
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Short/Abstract papers
Features exploratory work (4 pages or 2 pages excluding references) that may be preliminary but presents innovative concepts, early results, or thought-provoking viewpoints that stimulate discussion and future work.
Position papers
Offer critical perspectives on trends and challenges within the field (no less than 8 pages).
Survey papers
Provide thorough reviews of specific topics, mapping the current research landscape and suggesting directions for future exploration (no less than 8 pages).
Research areas in scope
01
Interactive Dimensions
What visual capability is needed for fully autonomous systemsHow can the vision community contribute to general-purpose robotic systemsWhat data modality is critical for generalizable, robust robot control
02
Stage-Wise Challenges
Data: Dynamic logs with multi-modal feedback. Teaching models about risk. Physically accurate data for the "sim-to-real" gap.Model: 3D geometry and physical dynamics. Differentiable cause-effect relationships. Conditioned representations over passive observation.Optimization: Safety and stability in the learning objective. Downstream task success. Model confidence for safe real-world deployment.Evaluation: Closed-loop performance. Reliability against environmental variability. Physical task completion and safe interaction.
Policies worth checking twice
- Submissions must follow CVPR two-column style and be anonymous.
- All formats allow unlimited references and appendices.
- Contributions are non-archival and hosted on the workshop website.
- Dual submission is allowed where permitted by third parties.
- Submissions under review or accepted by other conferences are welcome, and authors must mention this in the last sentence of the abstract.
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026