Towards credible visual model interpretation with path attribution

Naveed Akhtar, Mohammad A. A. K. Jalwana
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:439-457, 2023.

Abstract

With its inspirational roots in game-theory, path attribution framework stands out among the post-hoc model interpretation techniques due to its axiomatic nature. However, recent developments show that despite being axiomatic, path attribution methods can compute counter-intuitive feature attributions. Not only that, for deep visual models, the methods may also not conform to the original game-theoretic intuitions that are the basis of their axiomatic nature. To address these issues, we perform a systematic investigation of the path attribution framework. We first pinpoint the conditions in which the counter-intuitive attributions of deep visual models can be avoided under this framework. Then, we identify a mechanism of integrating the attributions over the paths such that they computationally conform to the original insights of game-theory. These insights are eventually combined into a method, which provides intuitive and reliable feature attributions. We also establish the findings empirically by evaluating the method on multiple datasets, models and evaluation metrics. Extensive experiments show a consistent quantitative and qualitative gain in the results over the baselines.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-akhtar23a, title = {Towards credible visual model interpretation with path attribution}, author = {Akhtar, Naveed and Jalwana, Mohammad A. A. K.}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {439--457}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/akhtar23a/akhtar23a.pdf}, url = {https://proceedings.mlr.press/v202/akhtar23a.html}, abstract = {With its inspirational roots in game-theory, path attribution framework stands out among the post-hoc model interpretation techniques due to its axiomatic nature. However, recent developments show that despite being axiomatic, path attribution methods can compute counter-intuitive feature attributions. Not only that, for deep visual models, the methods may also not conform to the original game-theoretic intuitions that are the basis of their axiomatic nature. To address these issues, we perform a systematic investigation of the path attribution framework. We first pinpoint the conditions in which the counter-intuitive attributions of deep visual models can be avoided under this framework. Then, we identify a mechanism of integrating the attributions over the paths such that they computationally conform to the original insights of game-theory. These insights are eventually combined into a method, which provides intuitive and reliable feature attributions. We also establish the findings empirically by evaluating the method on multiple datasets, models and evaluation metrics. Extensive experiments show a consistent quantitative and qualitative gain in the results over the baselines.} }
Endnote
%0 Conference Paper %T Towards credible visual model interpretation with path attribution %A Naveed Akhtar %A Mohammad A. A. K. Jalwana %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-akhtar23a %I PMLR %P 439--457 %U https://proceedings.mlr.press/v202/akhtar23a.html %V 202 %X With its inspirational roots in game-theory, path attribution framework stands out among the post-hoc model interpretation techniques due to its axiomatic nature. However, recent developments show that despite being axiomatic, path attribution methods can compute counter-intuitive feature attributions. Not only that, for deep visual models, the methods may also not conform to the original game-theoretic intuitions that are the basis of their axiomatic nature. To address these issues, we perform a systematic investigation of the path attribution framework. We first pinpoint the conditions in which the counter-intuitive attributions of deep visual models can be avoided under this framework. Then, we identify a mechanism of integrating the attributions over the paths such that they computationally conform to the original insights of game-theory. These insights are eventually combined into a method, which provides intuitive and reliable feature attributions. We also establish the findings empirically by evaluating the method on multiple datasets, models and evaluation metrics. Extensive experiments show a consistent quantitative and qualitative gain in the results over the baselines.
APA
Akhtar, N. & Jalwana, M.A.A.K.. (2023). Towards credible visual model interpretation with path attribution. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:439-457 Available from https://proceedings.mlr.press/v202/akhtar23a.html.

Related Material