Class-Incremental Learning with Cross-Space Clustering and Controlled Transfer

By Arjun Ashok, K. J. Joseph, Vineeth Nallure Balasubramanian. 2022

33 citations recorded by the source corpus.

In class-incremental learning, the model is expected to learn new classes continually while maintaining knowledge on previous classes. The challenge here lies in preserving the model's ability to effectively represent prior classes in the feature space, while adapting it to represent incoming new classes. We propose two distillation-based objectives for class incremental learning that leverage the structure of the feature space to maintain accuracy on previous classes, as well as enable learning the new classes. In our first objective, termed cross-space clustering (CSC), we propose to use the feature space structure of the previous model to characterize directions of optimization that maximally preserve the class: directions that all instances of a specific class should collectively optimize towards, and those that they should collectively optimize away from. Apart from minimizing forgetting, this indirectly encourages the model to cluster all instances of a class in the current feature space, and gives rise to a sense of herd-immunity, allowing all samples of a class to jointly combat the model from forgetting the class. Our second objective termed controlled transfer (CT) tackles incremental learning from an understudied perspective of inter-class transfer. CT explicitly approximates and conditions the current model on the semantic similarities between incrementally arriving classes and prior classes. This allows the model to learn classes in such a way that it maximizes positive forward transfer from similar prior classes, thus increasing plasticity, and minimizes negative backward transfer on dissimilar prior classes, whereby strengthening stability. We perform extensive experiments on two benchmark datasets, adding our method (CSCCT) on top of three prominent class-incremental learning methods. We observe consistent performance improvement on a variety of experimental settings.

Subfields: Artificial Intelligence, Cancer Research, Epidemiology, Meta & Continual Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Hang Su
2. Jun Zhu
3. Liyuan Wang
4. Xingxing Zhang
5. Matteo Boschini
6. Simone Calderara
7. Cosimo Fiorini
8. Riccardo Benaglia

In Cancer Research

1. Jun Zhu

In Epidemiology

1. Liyuan Wang
2. Jun Zhu
3. Hang Su
4. Xingxing Zhang

In Meta & Continual Learning

1. Xingxing Zhang
2. Liyuan Wang
3. Hang Su
4. Jun Zhu
5. Simone Calderara
6. Matteo Boschini
7. Cosimo Fiorini
8. Riccardo Benaglia

Citations in the corpus

• A Comprehensive Survey of Continual Learning: Theory, Method and Application (2023, methodology): “To adequately exploit the past information, DDE [152] distills colliding effects from the features of the new training samples, while CSC [153] additionally leverages the structure of the old feature space.”
• CaSpeR: Latent Spectral Regularization for Continual Learning (2023, methodology): “SCR [30] enforces consistency between two views of the input batch by leveraging the Supervised Contrastive loss [24]; differently, CSCCT [4] pairs an explicit latent-space clustering objec-tive that prevents the dispersion of previously-learned class boundaries with a controlled transfer objective…”