# Do Explanations make VQA Models more Predictable to a Human?

Source: [Do Explanations make VQA Models more Predictable to a Human?](<https://www.sota2.com/research/paper/68fa65001e209131cea07c24>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-27T05:56:48.736Z

Authors: Arjun Chandrasekaran, Viraj Prabhu, Deshraj Yadav, Prithvijit Chattopadhyay, Devi Parikh

Published: 2018-10-29T19:14:26.000Z

## Abstract

A rich line of research attempts to make deep neural networks more transparent by generating human-interpretable 'explanations' of their decision process, especially for interactive tasks like Visual Question Answering (VQA). In this work, we analyze if existing explanations indeed make a VQA model -- its responses as well as failures -- more predictable to a human. Surprisingly, we find that they do not. On the other hand, we find that human-in-the-loop approaches that treat the model as a black-box do.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/1810.12366>)

[PDF](<https://arxiv.org/pdf/1810.12366.pdf>)