Bridging Lifelong and Multi-Task Representation Learning via Algorithm and Complexity Measure

2025-11-03

AI Summary

This paper addresses lifelong learning under dynamic task arrival and online data acquisition, focusing on leveraging latent shared structure to accelerate learning of new tasks. We propose a general online algorithmic framework based on multi-task empirical risk minimization, requiring no prior knowledge of the full task sequence. A key innovation is the introduction of the “task-disentanglement dimension” as a unified complexity measure—first theoretically characterizing representation sharing in both lifelong and multi-task learning, and enabling statistical analysis for arbitrary function classes. We establish an upper bound on sample complexity that depends explicitly on this dimension, demonstrating significant improvement over naive baselines. Experiments confirm strong generalization and efficient knowledge transfer across noisy classification and regression tasks.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Search and Optimization: Learning to Search
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: The sharing economy
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

In lifelong learning, a learner faces a sequence of tasks with shared structure and aims to identify and leverage it to accelerate learning. We study the setting where such structure is captured by a common representation of data. Unlike multi-task learning or learning-to-learn, where tasks are available upfront to learn the representation, lifelong learning requires the learner to make use of its existing knowledge while continually gathering partial information in an online fashion. In this paper, we consider a generalized framework of lifelong representation learning. We propose a simple algorithm that uses multi-task empirical risk minimization as a subroutine and establish a sample complexity bound based on a new notion we introduce--the task-eluder dimension. Our result applies to a wide range of learning problems involving general function classes. As concrete examples, we instantiate our result on classification and regression tasks under noise.

Problem

• Bridging lifelong and multi-task learning through shared representation
• Developing algorithm for online learning with existing knowledge utilization
• Establishing sample complexity bounds using task-eluder dimension

Innovation

• Multi-task empirical risk minimization for lifelong learning
• Introduces task-eluder dimension for complexity analysis
• Applies to classification and regression under noise