• Home
  • Search
  • Reliable Visual Question Answering: Abstain Rather Than Answer Incorrectly
  • Open Access IconOpen Access
  • Cite Icon28
  • https://doi.org/10.1007/978-3-031-20059-5_9Copy DOI Icon

Reliable Visual Question Answering: Abstain Rather Than Answer Incorrectly

  • Jan 1, 2022
  • Spencer Whitehead +6 more
Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Abstract Machine learning has advanced dramatically, narrowing the accuracy gap to humans in multimodal tasks like visual question answering (VQA). However, while humans can say “I don’t know” when they are uncertain (i.e., abstain from answering a question), such ability has been largely neglected in multimodal research, despite the importance of this problem to the usage of VQA in real settings. In this work, we promote a problem formulation for reliable VQA, where we prefer abstention over providing an incorrect answer. We first enable abstention capabilities for several VQA models, and analyze both their coverage, the portion of questions answered, and risk, the error on that portion. For that, we explore several abstention approaches. We find that although the best performing models achieve over 71% accuracy on the VQA v2 dataset, introducing the option to abstain by directly using a model’s softmax scores limits them to answering less than 8% of the questions to achieve a low risk of error (i.e., 1%). This motivates us to utilize a multimodal selection function to directly estimate the correctness of the predicted answers, which we show can increase the coverage by, for example, \(2.4\times \) from 6.8% to 16.3% at 1% risk. While it is important to analyze both coverage and risk, these metrics have a trade-off which makes comparing VQA models challenging. To address this, we also propose an Effective Reliability metric for VQA that places a larger cost on incorrect answers compared to abstentions. This new problem formulation, metric, and analysis for VQA provide the groundwork for building effective and reliable VQA models that have the self-awareness to abstain if and only if they don’t know the answer. Code and Models: https://github.com/facebookresearch/reliable_vqa.

Similar Papers
  • PDF
  • Conference Article
  • Citations2

Can Pre-training help VQA with Lexical Variations?

  • Jan 01, 2020
  • Shailza Jolly +1
  • Research Article
  • Citations1

Adversarial Sample Synthesis for Visual Question Answering

  • Nov 21, 2024
  • ACM Transactions on Multimedia Computing, Communications, and Applications
  • Chuanhao Li +4
  • Conference Article
  • Citations668

Don't Just Assume; Look and Answer: Overcoming Priors for Visual Question Answering

  • Jun 01, 2018
  • Aishwarya Agrawal +3
  • Research Article
  • Citations42

Knowledge-Routed Visual Question Reasoning: Challenges for Deep Representation Embedding.

  • Jan 01, 2021
  • IEEE Transactions on Neural Networks and Learning Systems
  • Qingxing Cao +4
  • Research Article

Visual Question Answering Model Using Transformer-Based Multi-Head Cross Attention Network with Optimized Gold Rush Octave Convolution Network

  • Oct 28, 2025
  • International Journal of Image and Graphics
  • J Jinu Sophia +2
  • Conference Article
  • Citations47

Quantifying and Alleviating the Language Prior Problem in Visual Question Answering

  • Jul 18, 2019
  • Yangyang Guo +5
  • Research Article
  • Citations503

Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering

  • Sep 11, 2018
  • International Journal of Computer Vision
  • Yash Goyal +5
  • PDF
  • Research Article
  • Citations19

Advancing surgical VQA with scene graph knowledge

  • May 23, 2024
  • International Journal of Computer Assisted Radiology and Surgery
  • Kun Yuan +5
  • Book Chapter
  • Citations5

Neural Networks for Detecting Irrelevant Questions During Visual Question Answering

  • Jan 01, 2020
  • Mengdi Li +2
  • Conference Article
  • Citations7

Unleashing the Potential of Large Language Model: Zero-shot VQA for Flood Disaster Scenario

  • Nov 17, 2023
  • Yimin Sun +2
  • Research Article
  • Citations24

Enhancing robust VQA via contrastive and self-supervised learning

  • Nov 02, 2024
  • Pattern Recognition
  • Runlin Cao +4
  • Conference Article

Estimation of Visual Contents from Human Brain Signals via VQA Based on Brain-Specific Attention

  • Jun 04, 2023
  • Ryo Shichida +4
  • Research Article

VTFCGNet: a novel cross-modal reasoning network integrating Fourier self-attention and graph attention for visual text question answering

  • Feb 01, 2026
  • Neural Computing and Applications
  • Yujie Huo +3
  • Conference Article
  • Citations15

Feature Enhancement in Attention for Visual Question Answering

  • Jul 01, 2018
  • Yuetan Lin +3
  • Research Article
  • Citations30

Inverse Visual Question Answering: A New Benchmark and VQA Diagnosis Tool.

  • Nov 09, 2018
  • IEEE Transactions on Pattern Analysis and Machine Intelligence
  • Feng Liu +4
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.