Lossy compression of matrices by black-box optimisation of mixed integer nonlinear programming

TL;DR

This paper introduces a novel approach for lossy matrix compression in edge computing by combining black-box optimisation algorithms with an Ising solver, enabling efficient handling of mixed-integer nonlinear programming problems.

Contribution

It advances matrix compression techniques by integrating black-box optimisation with Ising solvers, improving optimization of mixed-integer nonlinear problems.

Findings

Ising solvers like simulated annealing, quantum annealing, and simulated quenching are compared.

The proposed method effectively optimizes matrix decompositions for lossy compression.

Different BBO strategies show varied performance, guiding future developments.

Abstract

In edge computing, suppressing data size is a challenge for machine learning models that perform complex tasks such as autonomous driving, in which computational resources (speed, memory size and power) are limited. Efficient lossy compression of matrix data has been introduced by decomposing it into the product of an integer and real matrices. However, its optimisation is difficult as it requires simultaneous optimisation of an integer and real variables. In this paper, we improve this optimisation by utilising recently developed black-box optimisation (BBO) algorithms with an Ising solver for integer variables. In addition, the algorithm can be used to solve mixed-integer programming problems that are linear and non-linear in terms of real and integer variables, respectively. The differences between the choice of Ising solvers (simulated annealing, quantum annealing and simulated quenching) and the strategies of the BBO algorithms (BOCS, FMQA and their variations) are discussed for further development of the BBO techniques.