Learning for VMM + WTA Embedded Classifiers
نویسندگان
چکیده
The authors present training and feedforward computation for a single layer of a VMM+WTA classifier. The experimental demonstration of the one-layer universal approximator encourages the use of one-layer networks for embedded low-power classification. The results enabling correct classification of each novel acoustic signal (generator, idle car, and idle truck). The classification structure requires, after training, less than 30μW of operational power and lower with additional fabrication.
منابع مشابه
VMM + WTA Embedded Classifiers Learning Algorithm implementable on SoC FPAA devices
This paper presents a learning algorithm for a VMM + WTA classifier one layer architecture on a Large-Scale Field Programmable Analog Array (FPAA). The technique enables opportunities for embedded, ultra-low power machine learning, techniques typically considered for large servers. To develop this training algorithm, the paper starts by understanding fundamental equivalent transformations for V...
متن کاملSoC FPAA Hardware Implementation of a VMM+WTA Embedded Learning Classifier
This paper focuses on the circuit aspects required for an on-chip, on-line SoC large-scale Field Programmable Analog Array (FPAA) learning for Vector-Matrix Multiplier (VMM) + Winner-Take-All (WTA) classifier structure. We start by describing the VMM+WTA classifier structure, and then show techniques required to handle device mismatch. The approach is initially explained using a VMM+WTA as a tw...
متن کاملClassifying with Gaussian Mixtures and Clusters
In this paper, we derive classifiers which are winner-take-all (WTA) approximations to a Bayes classifier with Gaussian mixtures for class conditional densities. The derived classifiers include clustering based algorithms like LVQ and k-Means. We propose a constrained rank Gaussian mixtures model and derive a WTA algorithm for it. Our experiments with two speech classification tasks indicate th...
متن کاملImproving Real-Time Performance of a Virtual Machine Monitor Based System
This paper describes our approach to enable Gandalf VMM (Virtual Machine Monitor) to be interruptible. Although Gandalf is shown to be a lightweight VMM, the detailed performance analysis using PMC (Performance Monitoring Counters) showed Gandalf executes with interrupts disabled for a rather long duration of time. By making Gandalf interruptible, we are able to make VMM based systems more suit...
متن کاملADT ADT ADT VMM Refinement memories ADT refinement ADTs
In this paper we present our novel power exploration methodology for applications with dynamic data types. Our methodology is crucial to obtain eeective solutions in an embedded (HW or SW) processor context. The contributions are twofold. First we deene the complete search space for Virtual Memory Management (VMM) mechanisms in a struc-tured way with orthogonal decision trees. Secondly we prese...
متن کامل