Self-Learning with Stochastic Triplet Loss

• Tipo: Artigo em Livro de Atas de Conferência Internacional
• Data: 2020

Autores

Pinto, JR (Autor) Outra; Jaime S Cardoso (Autor) FEUP

Ata de Conferência Internacional

Título: 2020 International Joint Conference on Neural Networks, IJCNN 2020, Glasgow, United Kingdom, July 19-24, 2020

Páginas: 1-8

Indexação

ISI Web of Knowledge - 2 Citações

Scopus - 3 Citações

Outras Informações

ID Authenticus: P-00S-RH1

DOI: 10.1109/ijcnn48605.2020.9206799

Abstract (EN): Deep learning has offered significant performance improvements on several pattern recognition problems. However, the well-known need for large amounts of labeled data limits applicability and performance where those are not available. Hence, this paper proposes an adaptation of the triplet loss for self-learning with entirely unlabeled data, where there is uncertainty in the generated triplets. The methodology was applied to off-the-person electrocardiogram-based biometric authentication and unconstrained face identity verification tasks, including stress experiments designed to simulate more difficult circumstances. Despite the uncertainty related to the use of unlabeled data, the method was mostly capable of avoiding negatively affecting the model's performance. The promising results show the proposed method can be a viable alternative to supervised learning in cases where only unlabeled data are available. The method is especially suitable for training with continuous stream-based datasets such as on person re-identification in video streams and continuous electrocardiogram-based biometrics.

Idioma: Inglês

Tipo (Avaliação Docente): Científica

Nº de páginas: 8