Friday, February 3, 2023
Latest:

Learning Center

Learn electronics, programming. inventing & MORE!!

Machine Learning 

tinyML Research Symposium 2021: Hypervector Design for Efficient Hyperdimensional Computing on…

tinyML 882 Comments , , , , , , ,



tinyML Research Symposium 2021 https://www.tinyml.org/event/research-symposium-2021
Hypervector Design for Efficient Hyperdimensional Computing on Edge Devices
Toygun BASAKLAR, PhD Student, University of Wisconsin-Madison

Hyperdimensional computing (HDC) has emerged as a new light-weight learning algorithm with smaller computation and energy requirements compared to conventional techniques. In HDC, data points are represented by high dimensional vectors (hypervectors), which are mapped to high dimensional space (hyperspace). Typically, a large hypervector dimension (= 1000) is required to achieve accuracies comparable to conventional alternatives. However, unnecessarily large hypervectors increase hardware and energy costs, which can undermine their benefits. This paper presents a technique to minimize the hypervector dimension while maintaining the accuracy and improving the robustness of the classifier. To this end, we formulate hypervector design as a multi-objective optimization problem for the first time in the literature. The proposed approach decreases the hypervector dimension by more than 128x while maintaining or increasing the accuracy achieved by conventional HDC. Experiments on a commercial hardware platform show that the proposed approach achieves more than two orders of magnitude reduction in model size, inference time, and energy consumption. We also demonstrate the trade-off between accuracy and robustness to noise and provide Pareto front solutions as a design parameter in our hypervector design.

source

You May Also Like

tinyML Talks: Software/Hardware Co-design for Tiny AI Systems

tinyML 347

Feature Selection for Knowledge Discovery and Data Mining (The Springer International Series in Engineering and Computer Science, 454)

admin 0

tinyML On Device Learning Forum – Valeria Tomaselli: Training models on tiny edge devices

tinyML 0

882 thoughts on “tinyML Research Symposium 2021: Hypervector Design for Efficient Hyperdimensional Computing on…”