See or Guess: Counterfactually Regularized Image Captioning

TL;DR

This paper introduces a causal inference-based framework for image captioning that enhances model robustness and interpretability, especially in counterfactual scenarios, by reducing hallucinations and improving faithfulness.

Contribution

It proposes a novel counterfactual regularization approach with two variants, improving generalizability and interpretability of image captioning models.

Findings

Reduces hallucinations in image captioning.

Enhances faithfulness to images across datasets.

Improves robustness in counterfactual scenarios.

Abstract

Image captioning, which generates natural language descriptions of the visual information in an image, is a crucial task in vision-language research. Previous models have typically addressed this task by aligning the generative capabilities of machines with human intelligence through statistical fitting of existing datasets. While effective for normal images, they may struggle to accurately describe those where certain parts of the image are obscured or edited, unlike humans who excel in such cases. These weaknesses they exhibit, including hallucinations and limited interpretability, often hinder performance in scenarios with shifted association patterns. In this paper, we present a generic image captioning framework that employs causal inference to make existing models more capable of interventional tasks, and counterfactually explainable. Our approach includes two variants leveraging either total effect or natural direct effect. Integrating them into the training process enables models to handle counterfactual scenarios, increasing their generalizability. Extensive experiments on various datasets show that our method effectively reduces hallucinations and improves the model's faithfulness to images, demonstrating high portability across both small-scale and large-scale image-to-text models. The code is available at https://github.com/Aman-4-Real/See-or-Guess.