Cancela, BraisBolón-Canedo, VerónicaAlonso-Betanzos, Amparo2024-11-202024-11-202021B. Cancela, V. Bolón-Canedo and A. Alonso-Betanzos, "Can data placement be effective for Neural Networks classification tasks? Introducing the Orthogonal Loss," 2020 25th International Conference on Pattern Recognition (ICPR), Milan, Italy, 2021, pp. 392-399, doi: 10.1109/ICPR48806.2021.9412704.http://hdl.handle.net/2183/40206Presented at: 5th International Conference on Pattern Recognition, ICPR 2020V, Virtual, Milan, 10-15 January 2021This version of the paper has been accepted for publication. 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. The final published paper is available online at: https://doi.org/10.1109/ICPR48806.2021.9412704[Abstract]: Traditionally, a Neural Network classification training loss function follows the same principle: minimizing the distance between samples that belong to the same class, while maximizing the distance to the other classes. There are no restrictions on the spatial placement of deep features (last layer input). This paper addresses this issue when dealing with Neural Networks, providing a set of loss functions that are able to train a classifier by forcing the deep features to be projected over a predefined orthogonal basis. Experimental results shows that these `data placement' functions can overcome the training accuracy provided by the classic cross-entropy loss function.eng© 2021 IEEE.TrainingNeural networksToolsPattern recognitionClassification algorithmsProposalsTask analysisconference outputopen access10.1109/ICPR48806.2021.9412704