Unleashing Hyperdimensional Computing with Nyström Method based Encoding
Keywords: Hyperdimensional Computing, Kernel Method
Abstract: Hyperdimensional computing (HDC) is an approach for solving cognitive information processing tasks using data represented as high dimensional, low-precision, vectors. The technique has a rigorous mathematical backing, and is easy to implement in energy-efficient and highly parallelizable hardware like FPGAs and ``in-memory'' architectures. The success of HDC based machine learning approaches is heavily dependent on the mapping from raw data to high-dimensional space. In this work, we propose a new method for constructing this mapping that is based on the Nyström method from the literature on kernel approximation. Our approach provides a simple recipe to turn any user-defined positive-semidefinite similarity function into an equivalent mapping in HDC. There is a vast literature on the design of such functions for learning problems, and our approach provides a mechanism to import them into the HDC setting, potentially expanding the types of problems that can be tackled using HDC. An empirical comparison of our approach against existing HDC encoding methods on a variety of classification tasks shows that we can achieve 10%-37% and 3%-18% better classification accuracy on graph and string datasets respectively.
Submission Number: 38
Loading