.png){kind=icon}
.png){kind=icon}
Return
Target-Oriented Autonomous Fuzzy Model Adaptation in Multimodal Transfer
.png){kind=icon}
PRE
AI
DOI：10.1109/TFUZZ.2026.3658508.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Fuzzy model-based domain adaptation has gained increasing attention for its ability to handle uncertainty arising from distribution shifts during knowledge transfer. However, existing fuzzy domain adaptation methods primarily focus on transferring information within the same data modality, leaving a gap in extending fuzzy domain adaptation to cross-modal scenarios. In addition, fuzzy domain adaptation across multiple domains still requires further exploration to effectively identify and select source models that are more relevant to the target domain. To address these gaps, this article proposes a target-oriented autonomous fuzzy model adaptation method built upon a pretrained multimodal foundation model, leveraging both visual and textual modalities. Simultaneously, an autonomous source selection strategy is developed by measuring similarities between each pair of source and target domains using fuzzy memberships, thereby enabling a multilayer fuzzy rule structure guided by the target domain. The proposed method is evaluated on three widely used public datasets, demonstrating the effectiveness of incorporating fuzzy rules and multimodal information.
Keywords:
Classification
domain adaptation
fuzzy rules
machine learning
transfer learning
Journal
IF:
11.9
Papers:
4.9K
Citations:
2.9W