Transferring Visual Explainability of Self-Explaining Models to Prediction-Only Models without Additional Training

2025/07/06 by Yuya Yoshikawa, Ryotaro Shimizu, Yoshikawa, Yuya +5 · 1 citation
Computer Science · #Artificial Intelligence (cs.AI) #Computer Vision and Pattern Recognition (cs.CV) #Data Visualization and Analytics #Explainable Artificial Intelligence (XAI) #FOS: Computer and information sciences #Machine Learning (cs.LG)

paper · pdf · doi:10.48550/arxiv.2507.04380

openalex publication_date 2025/07/06 · openalex created_date 2025/10/20 · openalex updated_date 2026/07/29

Abstract

In image classification scenarios where both prediction and explanation efficiency are required, self-explaining models that perform both tasks in a single inference are effective. However, for users who already have prediction-only models, training a new self-explaining model from scratch imposes significant costs in terms of both labeling and computation. This study proposes a method to transfer the visual explanation capability of self-explaining models learned in a source domain to prediction-only models in a target domain based on a task arithmetic framework. Our self-explaining model comprises an architecture that extends Vision Transformer-based prediction-only models, enabling the proposed method to endow explanation capability to many trained prediction-only models without additional training. Experiments on various image classification datasets demonstrate that, except for transfers between less-related domains, the transfer of visual explanation capability from source to target domains is successful, and explanation quality in the target domain improves without substantially sacrificing classification accuracy.

Citations

• DisTaC: Conditioning Task Vectors via Distillation for Robust Model Merging
• On Fairness of Task Arithmetic: The Role of Task Vectors
• Bring Reason to Vision: Understanding Perception and Reasoning through Model Merging
• Prediction via Shapley Value Regression
• A Comprehensive Survey on Self-Interpretable Neural Networks
• Bias Vector: Mitigating Biases in Language Models with Task Arithmetic Approach
• On the Surprising Effectiveness of Attention Transfer for Vision Transformers
• Selective Explanations
• TVE: Learning Meta-attribution for Transferable Vision Explainer
• Forgetting before Learning: Utilizing Parametric Arithmetic for Knowledge Updating in Large Language Models
• Saliency-Bench: A Comprehensive Benchmark for Evaluating Visual Explanations
• Strategies to exploit XAI to improve classification systems
• Efficient Shapley Values Estimation by Amortization for Text Classification
• DataComp: In search of the next generation of multimodal datasets
• Efficient XAI Techniques: A Taxonomic Survey
• Editing Models with Task Arithmetic
• Learning to Estimate Shapley Values with Vision Transformers
• Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time
• Explainable Artificial Intelligence for Autonomous Driving: A Comprehensive Overview and Field Guide for Future Research Directions
• Robust fine-tuning of zero-shot models
• Shapley Explanation Networks
• Learning Transferable Visual Models From Natural Language Supervision
• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
• Towards Robust Interpretability with Self-Explaining Neural Networks
• Neural Tangent Kernel: Convergence and Generalization in Neural Networks
• Averaging Weights Leads to Wider Optima and Better Generalization
• EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification
• A Unified Approach to Interpreting Model Predictions
• Right for the Right Reasons: Training Differentiable Models by\n Constraining their Explanations
• Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization
• Describing Textures in the Wild

Cited by

• DisTaC: Conditioning Task Vectors via Distillation for Robust Model Merging

Related

• Explainable Deep Classification Models for Domain Generalization
• Learning to Explain: A Model-Agnostic Framework for Explaining Black Box Models
• Sparks of Explainability: Recent Advancements in Explaining Large Vision Models
• TVE: Learning Meta-attribution for Transferable Vision Explainer
• Representing visual classification as a linear combination of words