Neuromorphic High Dimensional Computing Architecture for Classification Applications

High Dimensional Computing (HDC) includes the creation of High Dimensional Vectors, their storage in memory, and their functioning for different mathematical operations. The operations on HD vectors involve binding, bundling, and permutation, making it suitable to generate various kinds of architectures for performing different HDC applications. In this work, the High Dimensional Computing application involves language identification and classification of medical images using our proposed 2-D architectures. For the representation of the entities for the classification applications, 10000-bit random vectors are used. The high-dimensional entities are generated by using different random number generators and image processing techniques depending on the type of application. The experimental results, details about the data set used, architecture design information, benefits, and memory issues of HDC are also presented in this work. This new computational method carried out using high dimensional vectors introduces a new model for machine learning and can be used in different applications such as diagnosing Covid 19, brain tumour, classification of benign and malignant cancer cells, gesture classification and bio signal processing.