Abstract
dc:description.abstractDeep learning suffers from catastrophic forgetting when learning a dynamic environment, where the data is continuously gathered. This thesis is proposed to tackle the catastrophic forgetting problem under various settings and constraints, i.e., rehearsal-based, rehearsal-free, data scarcity, unlabeled data scarcity, non-i.i.d. distribution, and streaming data, applicable both in standalone and distributed systems. In addition, this thesis pushes the continual learning problem into 2 practical advanced sub-problems, i.e., unsupervised few-shot continual learning and federated few-shot class incremental learning. The proposed methods are evaluated in image data, including remote sensing data and time-series data.
Degree
thesis:*- College dc:contributor.school
- University of South Australia. UniSA STEM.
- Year dc:date.issued
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Ma’sum, Muhammad Anwar
Subjects
dc:subject × 5Rights
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/11541.2/44918