Abstract:Cross-Domain Few-Shot Learning (CD-FSL) addresses the Few-Shot Learning with a domain gap between source and target domains, which facilitates the transfer of knowledge from a source domain to a target domain with limited labeled samples. Current approaches often incorporate an auxiliary target dataset containing a few labeled samples to enhance model generalization on specific target domains. However, we observe that many models retain a substantial number of channels that learn source-specific knowledge and extract features that perform adequately on the source domain but generalize poorly to the target domain. This often results in compromised performance due to the influence of source-specific knowledge. To address this challenge, we introduce a novel framework, Gradient-Guided Channel Masking (GGCM), designed for CD-FSL to mitigate model channels from acquiring too much source-specific knowledge. GGCM quantifies each channel's contribution to solving target tasks using gradients of target loss and identifies those with smaller gradients as source-specific. These channels are then masked during the forward propagation of source features to mitigate the learning of source-specific knowledge. Conversely, GGCM mutes non-source-specific channels during the forward propagation of target features, forcing the model to depend on the source-specific channels and thereby enhancing their generalizability. Moreover, we propose a consistency loss that aligns the predictions made by source-specific channels with those made by the entire model. This approach further enhances the generalizability of these channels by enabling them to learn from the generalizable knowledge contained in other non-source-specific channels. Validated across multiple CD-FSL benchmark datasets, our framework demonstrates state-of-the-art performance and effectively suppresses the learning of source-specific knowledge.

Understanding Cross-Domain Few-Shot Learning Based on Domain Similarity and Few-Shot Difficulty

Enhancing Information Maximization with Distance-Aware Contrastive Learning for Source-Free Cross-Domain Few-Shot Learning

Cross-Domain Cross-Set Few-Shot Learning via Learning Compact and Aligned Representations

Deep Learning for Cross-Domain Few-Shot Visual Recognition: A Survey

Self-taught cross-domain few-shot learning with weakly supervised object localization and task-decomposition

How Well Do Self-Supervised Methods Perform in Cross-Domain Few-Shot Learning?

Cross-Domain Few-Shot Segmentation via Iterative Support-Query Correspondence Mining

Gradient-Guided Channel Masking for Cross-Domain Few-Shot Learning

Unified View Empirical Study for Large Pretrained Model on Cross-Domain Few-Shot Learning

ME-D2N: Multi-Expert Domain Decompositional Network for Cross-Domain Few-Shot Learning

TACDFSL: Task Adaptive Cross Domain Few-Shot Learning

Revisiting Mid-Level Patterns for Cross-Domain Few-Shot Recognition.

Soft Weight Pruning for Cross-Domain Few-Shot Learning with Unlabeled Target Data

Spectral Decomposition and Transformation for Cross-domain Few-shot Learning

Discriminativeness-Preserved Domain Adaptation for Few-Shot Learning

Domain-Adaptive Few-Shot Learning

Target Oriented Dynamic Adaption for Cross-Domain Few-Shot Learning

Cross-domain self-supervised few-shot learning via multiple crops with teacher-student network

Knowledge Transduction for Cross-Domain Few-Shot Learning

Cross-domain Self-supervised Learning for Domain Adaptation with Few Source Labels