IntentQA

IntentQA is a benchmark dataset used to evaluate human intention recognition in multimodal settings by assessing how well models can infer user goals from integrated text and visual cues.

Papers using IntentQA (3)

Intentvlm: Open-vocabulary Intention Recognition Through Forward-inverse Modeling With Video-language Models2026Progressive Video Condensation with MLLM Agent for Long-form Video Understanding2026BoxTuning: Directly Injecting the Object Box for Multimodal Model Fine-Tuning2026