Multiple Queries with Multiple Keys: A Precise Prompt Matching Paradigm for Prompt-based Continual Learning

2025-01-22

AI Summary

To address knowledge skew and catastrophic forgetting caused by inaccurate prompt selection in prompt-based continual learning, this paper proposes a Multi-Query Multi-Key (MQMK) matching paradigm. Our method introduces task-aware query vectors and a distribution-aware key encoder that jointly models class-prototype means and feature covariances—marking the first effort to explicitly align training data distributions as the core criterion for prompt selection. By synergistically combining broad-scale retrieval via multiple queries and fine-grained matching via multiple keys, MQMK achieves high-precision alignment between test samples and historical task distributions. Evaluated on three major continual learning benchmarks, MQMK establishes new state-of-the-art performance; under challenging scenarios, it improves prompt matching accuracy by over 30%. The implementation will be publicly released.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Search and Optimization: Learning to Search
• Data Mining & Knowledge Management: Intelligent Query Processing

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Continual learning requires machine learning models to continuously acquire new knowledge in dynamic environments while avoiding the forgetting of previous knowledge. Prompt-based continual learning methods effectively address the issue of catastrophic forgetting through prompt expansion and selection. However, existing approaches often suffer from low accuracy in prompt selection, which can result in the model receiving biased knowledge and making biased predictions. To address this issue, we propose the Multiple Queries with Multiple Keys (MQMK) prompt matching paradigm for precise prompt selection. The goal of MQMK is to select the prompts whose training data distribution most closely matches that of the test sample. Specifically, Multiple Queries enable precise breadth search by introducing task-specific knowledge, while Multiple Keys perform deep search by representing the feature distribution of training samples at a fine-grained level. Experiments show that MQMK enhances the prompt matching rate by over 30% in challenging scenarios and achieves state-of-the-art performance on three widely adopted continual learning benchmarks. Once this paper is accepted, we will release the code.

Problem

• Continual Learning
• Learning Bias
• Knowledge Retention

Innovation

• MQMK method
• Multiple Queries and Keys
• Prompt Matching Accuracy