Question Generation for Evaluating Cross-Dataset Shifts in Multi-modal Grounding

TL;DR

This paper introduces a question generation module that automatically creates out-of-distribution shifts to systematically evaluate how well VQA models adapt across different datasets, addressing domain mismatch issues.

Contribution

It proposes a novel VQG module for generating OOD shifts, enabling systematic evaluation of cross-dataset adaptation in VQA models.

Findings

The VQG module effectively generates diverse OOD shifts.

It helps identify specific domain mismatches affecting model performance.

The approach improves understanding of cross-dataset generalization in VQA.

Abstract

Visual question answering (VQA) is the multi-modal task of answering natural language questions about an input image. Through cross-dataset adaptation methods, it is possible to transfer knowledge from a source dataset with larger train samples to a target dataset where training set is limited. Suppose a VQA model trained on one dataset train set fails in adapting to another, it is hard to identify the underlying cause of domain mismatch as there could exists a multitude of reasons such as image distribution mismatch and question distribution mismatch. At UCLA, we are working on a VQG module that facilitate in automatically generating OOD shifts that aid in systematically evaluating cross-dataset adaptation capabilities of VQA models.