Rethinking Few-Shot Class-Incremental Learning with Open-Set Hypothesis in Hyperbolic Geometry
Cui, Yawen; Yu, Zitong; Peng, Wei; Tian, Qi; Liu, Li (2023-12-12)
IEEE
12.12.2023
Y. Cui, Z. Yu, W. Peng, Q. Tian and L. Liu, "Rethinking Few-Shot Class-Incremental Learning with Open-Set Hypothesis in Hyperbolic Geometry," in IEEE Transactions on Multimedia, doi: 10.1109/TMM.2023.3340550
https://rightsstatements.org/vocab/InC/1.0/
© 2023 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists,or reuse of any copyrighted component of this work in other works.
https://rightsstatements.org/vocab/InC/1.0/
© 2023 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists,or reuse of any copyrighted component of this work in other works.
https://rightsstatements.org/vocab/InC/1.0/
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:oulu-202403212376
https://urn.fi/URN:NBN:fi:oulu-202403212376
Tiivistelmä
Abstract
By training first with a large base dataset, FewShot Class-Incremental Learning (FSCIL) aims at continually learning a sequence of few-shot learning tasks with novel classes. There are mainly two challenges in FSCIL: the overfitting issue of novel classes with limited labeled samples and the catastrophic forgetting of previously seen classes. The current protocol of FSCIL is built by mimicking the general classincremental learning setting by building a unified framework, while the existing frameworks for FSCIL on this protocol always bias to the classes in the base dataset because the dominant performance of the deep model is decided by the size of the training dataset. Moreover, it is difficult to handle the stability-plasticity constraint in a unified FSCIL framework. To solve these issues, we rethink the configuration of FSCIL with the open-set hypothesis by reserving the possibility in the first session for incoming categories. To find a better decision boundary of close space and open space, Hyperbolic Reciprocal Point Learning module (Hyper-RPL) is built on Reciprocal Point Learning with hyperbolic neural networks. Besides, when learning novel categories from limited labeled data, we incorporate a hyperbolic metric learning (Hyper-Metric) module into the distillation-based framework to alleviate the overfitting issue and better handle the trade-off issue between the preservation of old knowledge and the acquisition of new knowledge. Finally, the comprehensive assessments of the proposed configuration and modules on three benchmark datasets are executed to validate the effectiveness, and state-of-the-art results are achieved.
By training first with a large base dataset, FewShot Class-Incremental Learning (FSCIL) aims at continually learning a sequence of few-shot learning tasks with novel classes. There are mainly two challenges in FSCIL: the overfitting issue of novel classes with limited labeled samples and the catastrophic forgetting of previously seen classes. The current protocol of FSCIL is built by mimicking the general classincremental learning setting by building a unified framework, while the existing frameworks for FSCIL on this protocol always bias to the classes in the base dataset because the dominant performance of the deep model is decided by the size of the training dataset. Moreover, it is difficult to handle the stability-plasticity constraint in a unified FSCIL framework. To solve these issues, we rethink the configuration of FSCIL with the open-set hypothesis by reserving the possibility in the first session for incoming categories. To find a better decision boundary of close space and open space, Hyperbolic Reciprocal Point Learning module (Hyper-RPL) is built on Reciprocal Point Learning with hyperbolic neural networks. Besides, when learning novel categories from limited labeled data, we incorporate a hyperbolic metric learning (Hyper-Metric) module into the distillation-based framework to alleviate the overfitting issue and better handle the trade-off issue between the preservation of old knowledge and the acquisition of new knowledge. Finally, the comprehensive assessments of the proposed configuration and modules on three benchmark datasets are executed to validate the effectiveness, and state-of-the-art results are achieved.
Kokoelmat
- Avoin saatavuus [38534]