Multimodal AI System
Design a system that interprets both text and image clues to answer quiz bowl questions, handling visual reasoning, OCR, diagram interpretation, and cross-modal fusion.
Machine Learning/CORE Unranked
EMM-QA 2026
ICML Workshop - Efficient Multimodal Question Answering
Jul 11 2026Official workshop site Site reachable
QANTA 2026 is the world’s first multimodal Quizbowl competition that challenges AI systems and human teams to answer questions combining text and visual clues such as photographs, diagrams, and artworks. It promotes human-AI collaboration and advances research in multimodal reasoning, with a parallel workshop at ICML 2026 inviting academic papers on efficient, incremental multimodal question answering.
Key deadlines
Verified September 9, 2026Full paper
May 30, 2026
AoE
Workshop timeline
Submission and decisions
Full paperKey deadline
May 30, 2026 · AoE
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Human Team Participation
Compete as a human player teaming up with an AI agent to answer multimodal questions and study human-AI complementarity.
Multimodal Question Author
Write pyramid-style tossup questions that integrate images with text clues, designed to be adversarial for AI while solvable by expert humans.
Research areas in scope
01
Multimodal Question Answering
Visual reasoningOCR and diagram interpretationCross-modal fusionAdversarial question designHuman-AI collaborationCalibration under incomplete cluesExplainability and robustness of multimodal systems
Policies worth checking twice
- Questions must be original ('clean') — not previously published
- No communication between question writers and computer competition teams
- Questions must follow PACE guidelines
- Images must be legally usable (public domain, CC-licensed, or fair use in an educational context)
- Images must relate directly to the answer and add genuine information beyond the text
- Text clues preceding an image must be resolvable without the image
- Text clues following an image must build on — not repeat — the image content
- All clues must be factually accurate and verifiable
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026