Local Path Integration for Attribution

Peiyu Yang, Naveed Akhtar, Zeyi Wen, Ajmal Mian

Research output: Chapter in Book/Conference paperConference paperpeer-review


Path attribution methods are a popular tool to interpret a visual model's prediction on an input. They integrate model gradients for the input features over a path defined between the input and a reference, thereby satisfying certain desirable theoretical properties. However, their reliability hinges on the choice of the reference. Moreover, they do not exhibit weak dependence on the input, which leads to counter-intuitive feature attribution mapping. We show that path-based attribution can account for the weak dependence property by choosing the reference from the local distribution of the input. We devise a method to identify the local input distribution and propose a technique to stochastically integrate the model gradients over the paths defined by the references sampled from that distribution. Our local path integration (LPI) method is found to consistently outperform existing path attribution techniques when evaluated on deep visual models. Contributing to the ongoing search of reliable evaluation metrics for the interpretation methods, we also introduce DiffID metric that uses the relative difference between insertion and deletion games to alleviate the distribution shift problem faced by existing metrics. Our code is available at https://github.com/ypeiyu/LPI.

Original languageEnglish
Title of host publicationAAAI-23 Technical Tracks 3
EditorsBrian Williams, Yiling Chen, Jennifer Neville
PublisherAAAI Press
Number of pages8
ISBN (Electronic)9781577358800
Publication statusPublished - 27 Jun 2023
Event37th AAAI Conference on Artificial Intelligence - Washington, United States
Duration: 7 Feb 202314 Feb 2023

Publication series

NameProceedings of the 37th AAAI Conference on Artificial Intelligence, AAAI 2023


Conference37th AAAI Conference on Artificial Intelligence
Abbreviated titleAAAI 2023
Country/TerritoryUnited States


Dive into the research topics of 'Local Path Integration for Attribution'. Together they form a unique fingerprint.

Cite this