Enhancing Model Interpretability with Local Attribution over Global Exploration

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhu, Zhiyu, Jin, Zhibo, Zhang, Jiayu, Chen, Huaming
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916358292242432
author Zhu, Zhiyu
Jin, Zhibo
Zhang, Jiayu
Chen, Huaming
author_facet Zhu, Zhiyu
Jin, Zhibo
Zhang, Jiayu
Chen, Huaming
contents In the field of artificial intelligence, AI models are frequently described as `black boxes' due to the obscurity of their internal mechanisms. It has ignited research interest on model interpretability, especially in attribution methods that offers precise explanations of model decisions. Current attribution algorithms typically evaluate the importance of each parameter by exploring the sample space. A large number of intermediate states are introduced during the exploration process, which may reach the model's Out-of-Distribution (OOD) space. Such intermediate states will impact the attribution results, making it challenging to grasp the relative importance of features. In this paper, we firstly define the local space and its relevant properties, and we propose the Local Attribution (LA) algorithm that leverages these properties. The LA algorithm comprises both targeted and untargeted exploration phases, which are designed to effectively generate intermediate states for attribution that thoroughly encompass the local space. Compared to the state-of-the-art attribution methods, our approach achieves an average improvement of 38.21\% in attribution effectiveness. Extensive ablation studies in our experiments also validate the significance of each component in our algorithm. Our code is available at: https://github.com/LMBTough/LA/
format Preprint
id arxiv_https___arxiv_org_abs_2408_07736
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Enhancing Model Interpretability with Local Attribution over Global Exploration
Zhu, Zhiyu
Jin, Zhibo
Zhang, Jiayu
Chen, Huaming
Machine Learning
Artificial Intelligence
In the field of artificial intelligence, AI models are frequently described as `black boxes' due to the obscurity of their internal mechanisms. It has ignited research interest on model interpretability, especially in attribution methods that offers precise explanations of model decisions. Current attribution algorithms typically evaluate the importance of each parameter by exploring the sample space. A large number of intermediate states are introduced during the exploration process, which may reach the model's Out-of-Distribution (OOD) space. Such intermediate states will impact the attribution results, making it challenging to grasp the relative importance of features. In this paper, we firstly define the local space and its relevant properties, and we propose the Local Attribution (LA) algorithm that leverages these properties. The LA algorithm comprises both targeted and untargeted exploration phases, which are designed to effectively generate intermediate states for attribution that thoroughly encompass the local space. Compared to the state-of-the-art attribution methods, our approach achieves an average improvement of 38.21\% in attribution effectiveness. Extensive ablation studies in our experiments also validate the significance of each component in our algorithm. Our code is available at: https://github.com/LMBTough/LA/
title Enhancing Model Interpretability with Local Attribution over Global Exploration
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2408.07736