EvQA

EvQA is the first objective, real-world benchmark for event-based multimodal large language models, consisting of 1,000 event-question-answer pairs sourced from 22 public datasets, used to evaluate the performance of these models in event-based visual question answering.

Papers using EvQA (2)

ProMSA:Progressive Multimodal Search Agents for Knowledge-Based Visual Question Answering2026Reconstruction as a Bridge for Event-Based Visual Question Answering2025