FRAMES-VQA: Benchmarking Fine-Tuning Robustness across Multi-Modal Shifts in Visual Question Answering

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Huang, Chengyue, Maneechotesuwan, Brisa, Chopra, Shivang, Kira, Zsolt
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916803307896832
author Huang, Chengyue
Maneechotesuwan, Brisa
Chopra, Shivang
Kira, Zsolt
author_facet Huang, Chengyue
Maneechotesuwan, Brisa
Chopra, Shivang
Kira, Zsolt
contents Visual question answering (VQA) systems face significant challenges when adapting to real-world data shifts, especially in multi-modal contexts. While robust fine-tuning strategies are essential for maintaining performance across in-distribution (ID) and out-of-distribution (OOD) scenarios, current evaluation settings are primarily unimodal or particular to some types of OOD, offering limited insight into the complexities of multi-modal contexts. In this work, we propose a new benchmark FRAMES-VQA (Fine-Tuning Robustness across Multi-Modal Shifts in VQA) for evaluating robust fine-tuning for VQA tasks. We utilize ten existing VQA benchmarks, including VQAv2, IV-VQA, VQA-CP, OK-VQA and others, and categorize them into ID, near and far OOD datasets covering uni-modal, multi-modal and adversarial distribution shifts. We first conduct a comprehensive comparison of existing robust fine-tuning methods. We then quantify the distribution shifts by calculating the Mahalanobis distance using uni-modal and multi-modal embeddings extracted from various models. Further, we perform an extensive analysis to explore the interactions between uni- and multi-modal shifts as well as modality importance for ID and OOD samples. These analyses offer valuable guidance on developing more robust fine-tuning methods to handle multi-modal distribution shifts. The code is available at https://github.com/chengyuehuang511/FRAMES-VQA .
format Preprint
id arxiv_https___arxiv_org_abs_2505_21755
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle FRAMES-VQA: Benchmarking Fine-Tuning Robustness across Multi-Modal Shifts in Visual Question Answering
Huang, Chengyue
Maneechotesuwan, Brisa
Chopra, Shivang
Kira, Zsolt
Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
Machine Learning
Visual question answering (VQA) systems face significant challenges when adapting to real-world data shifts, especially in multi-modal contexts. While robust fine-tuning strategies are essential for maintaining performance across in-distribution (ID) and out-of-distribution (OOD) scenarios, current evaluation settings are primarily unimodal or particular to some types of OOD, offering limited insight into the complexities of multi-modal contexts. In this work, we propose a new benchmark FRAMES-VQA (Fine-Tuning Robustness across Multi-Modal Shifts in VQA) for evaluating robust fine-tuning for VQA tasks. We utilize ten existing VQA benchmarks, including VQAv2, IV-VQA, VQA-CP, OK-VQA and others, and categorize them into ID, near and far OOD datasets covering uni-modal, multi-modal and adversarial distribution shifts. We first conduct a comprehensive comparison of existing robust fine-tuning methods. We then quantify the distribution shifts by calculating the Mahalanobis distance using uni-modal and multi-modal embeddings extracted from various models. Further, we perform an extensive analysis to explore the interactions between uni- and multi-modal shifts as well as modality importance for ID and OOD samples. These analyses offer valuable guidance on developing more robust fine-tuning methods to handle multi-modal distribution shifts. The code is available at https://github.com/chengyuehuang511/FRAMES-VQA .
title FRAMES-VQA: Benchmarking Fine-Tuning Robustness across Multi-Modal Shifts in Visual Question Answering
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
Machine Learning
url https://arxiv.org/abs/2505.21755