Task-conditioned Ensemble of Expert Models for Continuous Learning

2025-04-11

AI Summary

To address catastrophic forgetting in continual learning under non-stationary environments, this paper proposes a replay-free, fine-tuning-free task-conditioned ensemble of experts. The method tackles three key challenges: (1) it introduces an in-domain model grounded in local outlier detection to infer the task membership of incoming samples in real time; (2) it establishes a dynamic weighted ensemble mechanism that adaptively activates and fuses expert models based on runtime task attribution; and (3) it unifies modeling of cross-task, intra-task, and disjoint distributional shifts. Evaluated on LivDet-Iris-2017, LivDet-Iris-2020, and Split MNIST benchmarks, the approach significantly outperforms state-of-the-art continual learning methods, substantially mitigating accuracy degradation. The implementation is publicly available.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Learning to Search

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
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

One of the major challenges in machine learning is maintaining the accuracy of the deployed model (e.g., a classifier) in a non-stationary environment. The non-stationary environment results in distribution shifts and, consequently, a degradation in accuracy. Continuous learning of the deployed model with new data could be one remedy. However, the question arises as to how we should update the model with new training data so that it retains its accuracy on the old data while adapting to the new data. In this work, we propose a task-conditioned ensemble of models to maintain the performance of the existing model. The method involves an ensemble of expert models based on task membership information. The in-domain models-based on the local outlier concept (different from the expert models) provide task membership information dynamically at run-time to each probe sample. To evaluate the proposed method, we experiment with three setups: the first represents distribution shift between tasks (LivDet-Iris-2017), the second represents distribution shift both between and within tasks (LivDet-Iris-2020), and the third represents disjoint distribution between tasks (Split MNIST). The experiments highlight the benefits of the proposed method. The source code is available at https://github.com/iPRoBe-lab/Continuous_Learning_FE_DM.

Problem

• Maintain model accuracy in non-stationary environments
• Balance adaptation to new data while retaining old data accuracy
• Address distribution shifts between and within tasks

Innovation

• Task-conditioned ensemble of expert models
• Dynamic task membership via local outliers
• Maintains accuracy on old and new data