DualCP: Rehearsal-Free Domain-Incremental Learning via Dual-Level Concept Prototype

2025-03-23

AI Summary

This work addresses domain-incremental learning without data replay, tackling the core challenge of enabling visual models to continually adapt to sequential domain shifts while mitigating catastrophic forgetting. We propose a two-level concept prototype representation—comprising coarse-grained and fine-grained prototypes—alongside three novel components: a Concept Prototype Generator (CPG) for dynamic prototype construction, a Coarse-to-Fine Calibrator (C2F) for hierarchical alignment, and a Dual-Point Regression loss (DDR) for stable knowledge distillation. Our approach preserves historical knowledge and integrates new domain information efficiently—without storing past-domain samples or requiring full model retraining. Evaluated on DomainNet, CDDB, and CORe50 benchmarks, it consistently outperforms state-of-the-art replay-free methods, achieving average accuracy gains of 3.2%–5.7%. The framework significantly alleviates forgetting and enhances cross-domain generalization capability.

Technology Category

• Computer Vision: Representation Learning for Vision
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Domain-Incremental Learning (DIL) enables vision models to adapt to changing conditions in real-world environments while maintaining the knowledge acquired from previous domains. Given privacy concerns and training time, Rehearsal-Free DIL (RFDIL) is more practical. Inspired by the incremental cognitive process of the human brain, we design Dual-level Concept Prototypes (DualCP) for each class to address the conflict between learning new knowledge and retaining old knowledge in RFDIL. To construct DualCP, we propose a Concept Prototype Generator (CPG) that generates both coarse-grained and fine-grained prototypes for each class. Additionally, we introduce a Coarse-to-Fine calibrator (C2F) to align image features with DualCP. Finally, we propose a Dual Dot-Regression (DDR) loss function to optimize our C2F module. Extensive experiments on the DomainNet, CDDB, and CORe50 datasets demonstrate the effectiveness of our method.

Problem

• Addresses rehearsal-free domain-incremental learning challenges
• Resolves conflict between learning new and retaining old knowledge
• Enhances model adaptation to changing real-world conditions

Innovation

• Dual-level Concept Prototypes for class representation
• Concept Prototype Generator for coarse and fine prototypes
• Coarse-to-Fine calibrator aligns features with prototypes