# Explain and Improve: LRP-Inference Fine-Tuning for Image Captioning Models

arXiv: https://arxiv.org/abs/2001.01037 · HTML: https://mlsanity.com/paper/2001.01037/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 4 Jan 2020.

## Built on by (most cited first, 8 of 8)

- [Improving deep neural network generalization and robustness to background bias via layer-wise relevance propagation optimization](https://mlsanity.com/paper/2202.00232/) (2022)
- [Multimodal Explainable Artificial Intelligence: A Comprehensive Review of Methodological Advances and Future Research Directions](https://mlsanity.com/paper/2306.05731/) (2023)
- [A Survey on Explainable Artificial Intelligence for Cybersecurity](https://mlsanity.com/paper/2303.12942/) (2023)
- [Rethinking Explainability in the Era of Multimodal AI](https://mlsanity.com/paper/2506.13060/) (2025)
- [Long Exposure: Accelerating Parameter-Efficient Fine-Tuning for LLMs under Shadowy Sparsity](https://mlsanity.com/paper/2510.15964/) (2025)
- [Improving Reliability of Fine-tuning with Block-wise Optimisation](https://mlsanity.com/paper/2301.06133/) (2023)
- [Relevance-guided Neural Machine Translation](https://mlsanity.com/paper/2312.00214/) (2023)
- [Always Keep Your Promises: A Model-Agnostic Attribution Algorithm for Neural Networks](https://mlsanity.com/paper/2512.07010/) (2025)

## Abstract (arXiv)

This paper analyzes the predictions of image captioning models with attention mechanisms beyond visualizing the attention itself. We develop variants of layer-wise relevance propagation (LRP) and gradient-based explanation methods, tailored to image captioning models with attention mechanisms. We compare the interpretability of attention heatmaps systematically against the explanations provided by explanation methods such as LRP, Grad-CAM, and Guided Grad-CAM. We show that explanation methods provide simultaneously pixel-wise image explanations (supporting and opposing pixels of the input image) and linguistic explanations (supporting and opposing words of the preceding sequence) for each word in the predicted captions. We demonstrate with extensive experiments that explanation methods 1) can reveal additional evidence used by the model to make decisions compared to attention; 2) correlate to object locations with high precision; 3) are helpful to "debug" the model, e.g. by analyzing the reasons for hallucinated object words. With the observed properties of explanations, we further design an LRP-inference fine-tuning strategy that reduces the issue of object hallucination in image captioning models, and meanwhile, maintains the sentence fluency. We conduct experiments with two widely used attention mechanisms: the adaptive attention mechanism calculated with the additive attention and the multi-head attention mechanism calculated with the scaled dot product.