Continual Hyperbolic Learning of Instances and Classes
Abstract: Instance-level continual learning addresses the challenging problem of recognizing and remembering specific instances of object classes in an incremental setup, where new instances appear over time. Continual learning of instances forms a more fine-grained challenge than conventional continual learning, which is only concerned with incremental discrimination at the class level. In this paper, we argue that for real-world continual understanding, we need to recognize samples both at the instance and class level. We find that classes and instances form a hierarchical structure and propose HyperCLIC, a hyperbolic continual learning algorithm for visual instances and classes, to enable us to learn from this structure. We introduce continual hyperbolic classification and distillation objectives, allowing us to embed the hierarchical relations between classes and from classes to instances. Empirical evaluations show that HyperCLIC can operate effectively at both levels of granularity and with better hierarchical generalization, outperforming well-known continual learning algorithms. The code is included with this submission and will be made publicly available.
Submission Length: Regular submission (no more than 12 pages of main content)
Assigned Action Editor: ~Xuming_He3
Submission Number: 3158
Loading