Unlike exemplar-based class-incremental learning (EBCIL) which allows storing some old samples, exemplar-free class-incremental learning (EFCIL) faces a more severe forgetting problem due to the complete prohibition on accessing old data. Some previous methods freeze the feature extractor after the ...
Supplementary material for "FeTrIL: Feature Translation for Exemplar-Free Class-Incremental Learning" Gre´goire Petit1,2, Adrian Popescu1, Hugo Schindler1, David Picard2, Bertrand Delezoide3 1Universite´ Paris-Saclay, CEA, LIST, F-91120, Palaiseau, France 2LIGM, Ecole des Ponts, U...
1. Introduction Dynamic AI systems have a continual learning nature to learn new class data. They are expected to adapt to new classes while maintaining the knowledge of old classes, i.e., free from forgetting problems [31]. To evaluate this, the following protocol of ...
Exemplar-free class-incremental learning (CIL) poses several challenges since it prohibits the rehearsal of data from previous tasks and thus suffers from catastrophic forgetting. Recent approaches to incrementally learning the classifier by freezing the feature extractor after the first task have gained ...
In this paper, we proposed an exemplar-free approach -- Analytic Online Class Incremental Learning (AOCIL). Instead of back-propagation, we design the Analytic Classifier (AC) updated by recursive least square, cooperating with a frozen backbone. AOCIL simultaneously achieves high accuracy, low ...
SSUL: Semantic Segmentation with Unknown Label for Exemplar-based Class-Incremental Learning Sungmin Cha1,2*, Beomyoung Kim3*, YoungJoon Yoo2,3, Taesup Moon1 * Equal contribution 1 Department of Electrical and Computer Engineering, Seoul National University 2 NAVER AI Lab 3 Face, NAVER Clova...
Class-incremental learning (CIL) under an exemplar-free constraint has presented a significant challenge. Existing methods adhering to this constraint are prone to catastrophic forgetting, far more so than replay-based techniques that retain access to past samples. In this paper, to solve the ...
Exemplar-Free Deep Incremental Hashing for Efficient Image Retrieval Deep hashing techniques have been advanced by CNNs' semantic representations. However, existing incremental hashing methods rely on original data to mainta... S Jia,C Ma,B Li,... - International Conference on Intelligent Computing ...
The exemplar-free class incremental learning requires classification models to learn new class knowledge incrementally without retaining any old samples. Recently, the framework based on parallel one-class classifiers (POC), which trains a one-class classifier (OCC) independently for each category, has...
Exemplar-free class-incremental learning (EFCIL) presents a significant challenge as the old class samples are absent for new task learning. Due to the severe imbalance between old and new class samples, the learned classifiers can be easily biased toward the new ones. Moreover, continually ...