HyCal: A Training-Free Prototype Calibration Method for Cross-Discipline Few-Shot Class-Incremental Learning

TL;DR

HyCal is a training-free prototype calibration method that improves cross-discipline few-shot class-incremental learning by addressing domain imbalance and heterogeneity in CLIP embeddings.

Contribution

The paper introduces HyCal, a novel training-free calibration technique combining cosine similarity and Mahalanobis distance to enhance prototype stability in heterogeneous, imbalanced domains.

Findings

HyCal outperforms existing methods in imbalanced cross-domain incremental learning.

HyCal effectively mitigates the effects of Domain Gravity in heterogeneous data.

Experiments demonstrate consistent improvements in prototype stability and performance.

Abstract

Pretrained Vision-Language Models (VLMs) like CLIP show promise in continual learning, but existing Few-Shot Class-Incremental Learning (FSCIL) methods assume homogeneous domains and balanced data distributions, limiting real-world applicability where data arises from heterogeneous disciplines with imbalanced sample availability and varying visual complexity. We identify Domain Gravity, a representational asymmetry where data imbalance across heterogeneous domains causes overrepresented or low-entropy domains to disproportionately influence the embedding space, leading to prototype drift and degraded performance on underrepresented or high-entropy domains. To address this, we introduce Cross-Discipline Variable Few-Shot Class-Incremental Learning (XD-VSCIL), a benchmark capturing real-world heterogeneity and imbalance where Domain Gravity naturally intensifies. We propose Hybrid Prototype Calibration (HyCal), a training-free method combining cosine similarity and Mahalanobis distance to capture complementary geometric properties-directional alignment and covariance-aware magnitude-yielding stable prototypes under imbalanced heterogeneous conditions. Operating on frozen CLIP embeddings, HyCal achieves consistent retention-adaptation improvements while maintaining efficiency. Experiments show HyCal effectively mitigates Domain Gravity and outperforms existing methods in imbalanced cross-domain incremental learning.