Weightless neural networks for open set recognition

Abstract

Open set recognition is a classification-like task. It is accomplished not only by the identification of observations which belong to targeted classes (i.e., the classes among those represented in the training sample which should be later recognized) but also by the rejection of inputs from other classes in the problem domain. The need for proper handling of elements of classes beyond those of interest is frequently ignored, even in works found in the literature. This leads to the improper development of learning systems, which may obtain misleading results when evaluated in their test beds, consequently failing to keep the performance level while facing some real challenge. The adaptation of a classifier for open set recognition is not always possible: the probabilistic premises most of them are built upon are not valid in a openset setting. Still, this paper details how this was realized for WiSARD a weightless artificial neural network model. Such achievement was based on an elaborate distance-like computation this model provides and the definition of rejection thresholds during training. The pro-Editors: Thomas Gärtner, Mirco Nanni, Andrea Passerini, and Celine Robardet. Douglas O. Cardoso thanks CAPES (process 99999.005992/2014-01) and CNPq for financial support. João Gama thanks the support of the European Commission through the project MAESTRA (Grant Number ICT-750 2013-612944). Felipe M. G. França thanks the support of FAPERJ, FINEP and INOVAX.

Recent Advances in Open Set Recognition: A Survey

“…Next, we first give a review from the discriminative model perspective, where most existing OSR algorithms are modeled from this perspective. Deep Neural Network-based [25], [64], [79]- [87] Generative model Instance Generation-based [62], [88]- [91] Non-Instance Generation-based [63] A. Discriminative Model for OSR 1) Traditional ML Methods-based OSR Models: As mentioned above, traditional machine learning methods (e.g., SVM, sparse representation, Nearest Neighbor, etc.) usually assume that the training and testing data are drawn from the same distribution.…”

“…The first SVM is a one-class SVM CAP model used as a conditioner: if the posterior estimate P O (y|x) of an input sample x predicted by one-class SVM is less than a threshold δ τ , the sample will be rejected outright. Otherwise, OSR Discriminative Model (DM) SVM-based [21], [65]- [69] + EVT TML-based [22], [23] SR-based [61] + EVT Dis-based [70], [71] MD-based [72], [73] + EVT Others [74], [75], [76] DNN-based [25], [64], [79], [81]- [84] + EVT [77], [78], [80], [85] Generative Model (GM)…”

“…Kardan and Stanley [113] proposed the competitive overcomplete output layer (COOL) neural network to circumvent the overgeneralization of neural networks over regions far from the training data. Based on an elaborate distance-like computation provided by a weightless neural network, Cardoso et al [84] proposed the tWiSARD algorithm for open set recognition, which is further developed in [64]. Recently, considering the available background classes (KUCs), Dhamija et al [25] combined SoftMax with the novel Entropic Open-Set and Objectosphere losses to address the OSR problem.…”

Can Reverse Nearest Neighbors Perceive Unknowns?

“…It separates plant images from unknown non-plant images. Open set recognition through weightless neural network has been explored in [8]. In [9], a neural network based classifier detects the unknown samples through comparison and computation of the similarity between the unknown data and the stored or bounded knowledge.…”

3-Step flow focusing enables multidirectional imaging of bioparticles for imaging flow cytometry

“…Thus, the reduction of false positive classifications and the identification of foreign particles remains a challenging task. 33 We demonstrate how such foreign particles can be reliably identified by our extended algorithm. As a proof of concept we show the suitability of our mIFC image series for 3D-reconstruction, demonstrated on human blood cells and pollen.…”