作者位次 #6/7ACL 2026Poster· Natural Language Processing
Looking Beyond the One: Operationalizing and Eliciting Visual Ambiguity in VLLM s
Yuchong Chen · Bowei Zou · Yuhan Chen · Yifan Fan · Xinyu Li · Shujun Cao · Yu Hong
摘要
Visual questions are often ambiguous: the same image–question pair may admit multiple valid answers depending on which region is referenced. However, current Visual Question Answering (VQA) systems typically collapse this ambiguity, committing to a single interpretation during decoding and evalua...