Language Guided Concept Bottleneck Models for Interpretable Continual Learning

2025-03-30

AI Summary

Continual learning (CL) faces the dual challenge of mitigating catastrophic forgetting while ensuring decision interpretability. To address this, we propose the first language-guided concept bottleneck model for CL, introducing a CLIP-based semantic alignment framework at the concept layer that establishes human-understandable and cross-task generalizable decision mechanisms. Our method enables differentiable, reusable concept-level reasoning via natural language supervision and supports concept-level attribution visualization, thereby jointly enhancing knowledge retention and decision transparency. Evaluated on the ImageNet-subset continual learning benchmark, our approach achieves a 3.06% improvement in average accuracy over state-of-the-art methods. By unifying semantic grounding, concept reuse, and interpretable inference within a continual learning setting, this work establishes a novel paradigm for explainable continual learning.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Computer Vision: Interpretability, Explainability, and Transparency
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Continual learning (CL) aims to enable learning systems to acquire new knowledge constantly without forgetting previously learned information. CL faces the challenge of mitigating catastrophic forgetting while maintaining interpretability across tasks. Most existing CL methods focus primarily on preserving learned knowledge to improve model performance. However, as new information is introduced, the interpretability of the learning process becomes crucial for understanding the evolving decision-making process, yet it is rarely explored. In this paper, we introduce a novel framework that integrates language-guided Concept Bottleneck Models (CBMs) to address both challenges. Our approach leverages the Concept Bottleneck Layer, aligning semantic consistency with CLIP models to learn human-understandable concepts that can generalize across tasks. By focusing on interpretable concepts, our method not only enhances the models ability to retain knowledge over time but also provides transparent decision-making insights. We demonstrate the effectiveness of our approach by achieving superior performance on several datasets, outperforming state-of-the-art methods with an improvement of up to 3.06% in final average accuracy on ImageNet-subset. Additionally, we offer concept visualizations for model predictions, further advancing the understanding of interpretable continual learning.

Problem

• Mitigate catastrophic forgetting in continual learning
• Maintain interpretability across evolving tasks
• Align semantic concepts for transparent decision-making

Innovation

• Integrates language-guided Concept Bottleneck Models
• Aligns semantic consistency with CLIP models
• Enhances interpretability and knowledge retention