Class Mean Vectors, Self Monitoring and Self Learning for Neural Classifiers
This addresses classification challenges for neural networks by offering a simple, widely applicable method, though it appears incremental as it builds on existing mean-based techniques.
The paper tackles the problem of classification by using sample mean vectors to compute neural network weights without training, monitor performance on unlabeled data, and enable self-training, achieving promising results on CIFAR-10.
In this paper we explore the role of sample mean in building a neural network for classification. This role is surprisingly extensive and includes: direct computation of weights without training, performance monitoring for samples without known classification, and self-training for unlabeled data. Experimental computation on a CIFAR-10 data set provides promising empirical evidence on the efficacy of a simple and widely applicable approach to some difficult problems.