A Survey of Post-Hoc XAI Methods From a Visualization Perspective: Challenges and Opportunities
XAI (eXplainable AI) has become a pivotal area of research with the advancement of deep learning (DL) technologies and applications. Post-hoc explanation methods interpret deep learning predictions by uncovering the significance of input features, while visualization tools can contribute to a deep u...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/11039632/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849317928637300736 |
|---|---|
| author | Deepshikha Bhati MD Amiruzzaman Ye Zhao Angela Guercio Tram Le |
| author_facet | Deepshikha Bhati MD Amiruzzaman Ye Zhao Angela Guercio Tram Le |
| author_sort | Deepshikha Bhati |
| collection | DOAJ |
| description | XAI (eXplainable AI) has become a pivotal area of research with the advancement of deep learning (DL) technologies and applications. Post-hoc explanation methods interpret deep learning predictions by uncovering the significance of input features, while visualization tools can contribute to a deep understanding of AI model reasoning based on these methods. In this paper, we survey a broad spectrum of post-hoc explanation methods and the visual analytics work based on them. First, we categorize the computational methods into four main types: perturbation-based, gradient-based, decomposition-based, and concept-based. While the first three focus on attributing the model’s output to specific regions of the input image, concept-based methods provide global explanations by mapping human-understandable concepts to high-level features. Then, we examine the methodologies, features, strengths, and limitations of each approach. Moreover, we review existing visualization-focused work based on these computational methods. Finally, we discuss further research challenges and opportunities for XAI visualization with post-hoc explanation. |
| format | Article |
| id | doaj-art-dd8fe3aa6f214e8ebaaa146926379986 |
| institution | Kabale University |
| issn | 2169-3536 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-dd8fe3aa6f214e8ebaaa1469263799862025-08-20T03:51:03ZengIEEEIEEE Access2169-35362025-01-011312078512080610.1109/ACCESS.2025.358113611039632A Survey of Post-Hoc XAI Methods From a Visualization Perspective: Challenges and OpportunitiesDeepshikha Bhati0https://orcid.org/0009-0002-0115-6026MD Amiruzzaman1https://orcid.org/0000-0002-2292-5798Ye Zhao2https://orcid.org/0000-0003-3877-0943Angela Guercio3Tram Le4https://orcid.org/0009-0002-6679-2584Department of Computer Science, Kent State University, Kent, OH, USADepartment of Computer Science, West Chester University, West Chester, PA, USADepartment of Computer Science, Kent State University, Kent, OH, USADepartment of Computer Science, Kent State University, Kent, OH, USADepartment of Computer Science, Kent State University, Kent, OH, USAXAI (eXplainable AI) has become a pivotal area of research with the advancement of deep learning (DL) technologies and applications. Post-hoc explanation methods interpret deep learning predictions by uncovering the significance of input features, while visualization tools can contribute to a deep understanding of AI model reasoning based on these methods. In this paper, we survey a broad spectrum of post-hoc explanation methods and the visual analytics work based on them. First, we categorize the computational methods into four main types: perturbation-based, gradient-based, decomposition-based, and concept-based. While the first three focus on attributing the model’s output to specific regions of the input image, concept-based methods provide global explanations by mapping human-understandable concepts to high-level features. Then, we examine the methodologies, features, strengths, and limitations of each approach. Moreover, we review existing visualization-focused work based on these computational methods. Finally, we discuss further research challenges and opportunities for XAI visualization with post-hoc explanation.https://ieeexplore.ieee.org/document/11039632/Deep learning visualizationexplainable AI (XAI)post-hoc explanationvisual analytics |
| spellingShingle | Deepshikha Bhati MD Amiruzzaman Ye Zhao Angela Guercio Tram Le A Survey of Post-Hoc XAI Methods From a Visualization Perspective: Challenges and Opportunities IEEE Access Deep learning visualization explainable AI (XAI) post-hoc explanation visual analytics |
| title | A Survey of Post-Hoc XAI Methods From a Visualization Perspective: Challenges and Opportunities |
| title_full | A Survey of Post-Hoc XAI Methods From a Visualization Perspective: Challenges and Opportunities |
| title_fullStr | A Survey of Post-Hoc XAI Methods From a Visualization Perspective: Challenges and Opportunities |
| title_full_unstemmed | A Survey of Post-Hoc XAI Methods From a Visualization Perspective: Challenges and Opportunities |
| title_short | A Survey of Post-Hoc XAI Methods From a Visualization Perspective: Challenges and Opportunities |
| title_sort | survey of post hoc xai methods from a visualization perspective challenges and opportunities |
| topic | Deep learning visualization explainable AI (XAI) post-hoc explanation visual analytics |
| url | https://ieeexplore.ieee.org/document/11039632/ |
| work_keys_str_mv | AT deepshikhabhati asurveyofposthocxaimethodsfromavisualizationperspectivechallengesandopportunities AT mdamiruzzaman asurveyofposthocxaimethodsfromavisualizationperspectivechallengesandopportunities AT yezhao asurveyofposthocxaimethodsfromavisualizationperspectivechallengesandopportunities AT angelaguercio asurveyofposthocxaimethodsfromavisualizationperspectivechallengesandopportunities AT tramle asurveyofposthocxaimethodsfromavisualizationperspectivechallengesandopportunities AT deepshikhabhati surveyofposthocxaimethodsfromavisualizationperspectivechallengesandopportunities AT mdamiruzzaman surveyofposthocxaimethodsfromavisualizationperspectivechallengesandopportunities AT yezhao surveyofposthocxaimethodsfromavisualizationperspectivechallengesandopportunities AT angelaguercio surveyofposthocxaimethodsfromavisualizationperspectivechallengesandopportunities AT tramle surveyofposthocxaimethodsfromavisualizationperspectivechallengesandopportunities |