Multi-View Class Incremental Learning
This work addresses the practical need for multi-view learning methods to operate in open-ended environments, though it appears incremental as it builds on existing incremental learning and multi-view techniques.
The paper tackles the problem of multi-view class incremental learning (MVCIL), where a model must incrementally classify new classes from a continual stream of views without access to earlier data, and addresses challenges like catastrophic forgetting and interference by developing a randomization-based representation learning technique, orthogonality fusion, and selective weight consolidation, achieving validated effectiveness in experiments.
Multi-view learning (MVL) has gained great success in integrating information from multiple perspectives of a dataset to improve downstream task performance. To make MVL methods more practical in an open-ended environment, this paper investigates a novel paradigm called multi-view class incremental learning (MVCIL), where a single model incrementally classifies new classes from a continual stream of views, requiring no access to earlier views of data. However, MVCIL is challenged by the catastrophic forgetting of old information and the interference with learning new concepts. To address this, we first develop a randomization-based representation learning technique serving for feature extraction to guarantee their separate view-optimal working states, during which multiple views belonging to a class are presented sequentially; Then, we integrate them one by one in the orthogonality fusion subspace spanned by the extracted features; Finally, we introduce selective weight consolidation for learning-without-forgetting decision-making while encountering new classes. Extensive experiments on synthetic and real-world datasets validate the effectiveness of our approach.