Lifelong Infinite Mixture Model Based on Knowledge-Driven Dirichlet Process

TL;DR

This paper introduces a theoretical framework and a novel Lifelong Infinite Mixture (LIMix) model for lifelong learning, which adaptively expands and manages knowledge to learn new tasks while avoiding catastrophic forgetting.

Contribution

It provides the first theoretical risk bounds for lifelong learning models and proposes a new LIMix approach using Dirichlet processes for adaptive knowledge integration.

Findings

Derived risk bounds based on discrepancy distance.

Developed the LIMix model with adaptive architecture expansion.

Trained a compact Student model for cross-domain inference.

Abstract

Recent research efforts in lifelong learning propose to grow a mixture of models to adapt to an increasing number of tasks. The proposed methodology shows promising results in overcoming catastrophic forgetting. However, the theory behind these successful models is still not well understood. In this paper, we perform the theoretical analysis for lifelong learning models by deriving the risk bounds based on the discrepancy distance between the probabilistic representation of data generated by the model and that corresponding to the target dataset. Inspired by the theoretical analysis, we introduce a new lifelong learning approach, namely the Lifelong Infinite Mixture (LIMix) model, which can automatically expand its network architectures or choose an appropriate component to adapt its parameters for learning a new task, while preserving its previously learnt information. We propose to incorporate the knowledge by means of Dirichlet processes by using a gating mechanism which computes the dependence between the knowledge learnt previously and stored in each component, and a new set of data. Besides, we train a compact Student model which can accumulate cross-domain representations over time and make quick inferences. The code is available at https://github.com/dtuzi123/Lifelong-infinite-mixture-model.