Skip to main navigation Skip to search Skip to main content

Attention Lattice Adapter: Visual Explanation Generation for Visual Foundation Models

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this study, we consider the problem of generating visual explanations in visual foundation models. Numerous methods have been proposed for this purpose; however, they often cannot be applied to complex models due to their lack of adaptability. To overcome these limitations, we propose a novel explanation generation method in visual foundation models that is aimed at both generating explanations and partially updating model parameters to enhance interpretability. Our approach introduces two novel mechanisms: Attention Lattice Adapter (ALA) and Alternating Epoch Architect (AEA). ALA mechanism simplifies the process by eliminating the need for manual layer selection, thus enhancing the model’s adaptability and interpretability. Moreover, the AEA mechanism, which updates ALA’s parameters every other epoch, effectively addresses the common issue of overly small attention regions. We evaluated our method on two benchmark datasets, CUB-200-2011 and ImageNet-S. Our results showed that our method outperformed the baseline methods in terms of mean intersection over union (IoU), insertion score, deletion score, and insertion-deletion score on both the CUB-200-2011 and ImageNet-S datasets. Notably, our best model achieved a 53.2-point improvement in mean IoU on the CUB-200-2011 dataset compared with the baselines.

Original languageEnglish
Title of host publicationNeural Information Processing - 32nd International Conference, ICONIP 2025, Proceedings
EditorsTadahiro Taniguchi, Chi Sing Andrew Leung, Tadashi Kozuno, Junichiro Yoshimoto, Mufti Mahmud, Maryam Doborjeh, Kenji Doya
PublisherSpringer Science and Business Media Deutschland GmbH
Pages395-410
Number of pages16
ISBN (Print)9789819540969
DOIs
Publication statusPublished - 2026
Event32nd International Conference on Neural Information Processing, ICONIP 2025 - Okinawa, Japan
Duration: 2025 Nov 202025 Nov 24

Publication series

NameCommunications in Computer and Information Science
Volume2756 CCIS
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference32nd International Conference on Neural Information Processing, ICONIP 2025
Country/TerritoryJapan
CityOkinawa
Period25/11/2025/11/24

Keywords

  • CLIP
  • Explainable AI
  • Interpretability

ASJC Scopus subject areas

  • General Computer Science
  • General Mathematics

Fingerprint

Dive into the research topics of 'Attention Lattice Adapter: Visual Explanation Generation for Visual Foundation Models'. Together they form a unique fingerprint.

Cite this