Few-Shot Domain Incremental Learning via Continual Vision-Language Consolidation

2026-06-29

AI Summary

This work addresses the challenges of overfitting and poor adaptation to new domains in few-shot domain incremental learning by formally defining the task and proposing a continual vision-language fusion approach. The method leverages large language models to generate multi-template textual prototypes, which are integrated with visual prototypes to construct cross-modal knowledge. To preserve knowledge from base domains, a latent space retention mechanism is introduced, and a dual-cluster projection scheme enables parameter-efficient fine-tuning. By synergistically combining shared and domain-specific components, the proposed framework substantially outperforms existing methods across multiple benchmarks, achieving performance gains of up to 16%. The implementation code has been made publicly available.

Abstract

Existing domain-incremental learning (DIL) strategies call for massive amounts of data to adapt to new domains and suffer from the overfitting problem in the case of data scarcity. This paper puts forward a relatively uncharted problem, namely, few-shot domain incremental learning (FSDIL), taking into account the problem of extreme data shortages in the realm of DIL. A novel algorithm, namely Continual Vision-Language Consolidation (CVLC), is proposed to address the FSDIL problem, where the key idea lies in the concept of latent space reservation in the base domain coupled with dual coalescent projection (DCP) as a parameter-efficient fine-tuning method. First, the vision prototype is calibrated while multiple templates and synonyms are generated via LLMs to induce the language prototype. The vision and language prototypes are fused. Adaptation to never-ending arrivals of new domains is done by the DCP technique, fine-tuned in such a way to prepare the model to unseen domains via latent-space reservations committed in the base domain. CVLC is structured under shared and domain-specific components to combine general knowledge and domain-specific details. The advantage of our approach is demonstrated through a range of benchmark problems and comparisons with prior arts, in which CVLC outperforms them by up to a 16% gap. Our codes are shared publicly in https://github.com/Naeem-Paeedeh/CVLC .

Problem

• Few-Shot Domain Incremental Learning
• Data Scarcity
• Domain Incremental Learning
• Overfitting

Innovation

• Few-Shot Domain Incremental Learning
• Continual Vision-Language Consolidation
• Latent Space Reservation
• Dual Coalescent Projection
• Parameter-Efficient Fine-Tuning