MFABA: A More Faithful and Accelerated Boundary-based Attribution Method for Deep Neural Networks

December 21, 2023 · Entered Twilight · 🏛 AAAI Conference on Artificial Intelligence

Repo contents: .DS_Store, .gitignore, README.md, appendix, eval.py, eval_all.sh, example.ipynb, fonts, generate_all.sh, generate_attributions.py, images, requirements.txt, saliency, utils.py, visualization.py

Authors Zhiyu Zhu, Huaming Chen, Jiayu Zhang, Xinyi Wang, Zhibo Jin, Minhui Xue, Dongxiao Zhu, Kim-Kwang Raymond Choo arXiv ID 2312.13630 Category cs.CV: Computer Vision Cross-listed cs.LG Citations 14 Venue AAAI Conference on Artificial Intelligence Repository https://github.com/LMBTough/MFABA ⭐ 2 Last Checked 1 month ago

Abstract

To better understand the output of deep neural networks (DNN), attribution based methods have been an important approach for model interpretability, which assign a score for each input dimension to indicate its importance towards the model outcome. Notably, the attribution methods use the axioms of sensitivity and implementation invariance to ensure the validity and reliability of attribution results. Yet, the existing attribution methods present challenges for effective interpretation and efficient computation. In this work, we introduce MFABA, an attribution algorithm that adheres to axioms, as a novel method for interpreting DNN. Additionally, we provide the theoretical proof and in-depth analysis for MFABA algorithm, and conduct a large scale experiment. The results demonstrate its superiority by achieving over 101.5142 times faster speed than the state-of-the-art attribution algorithms. The effectiveness of MFABA is thoroughly evaluated through the statistical analysis in comparison to other methods, and the full implementation package is open-source at: https://github.com/LMBTough/MFABA