Customized Data Compression: Automatically Synthesizing Effective Data Compression and Decompression Algorithms

Abstract

With the exponential increase in the amount of data humans are generating, there is a progressive need for developing effective high-speed data compression techniques. However, developing an algorithm that just compresses data well is not sufficient. A good compression technique should also minimize the compression and decompression time. This project focuses on automatically synthesizing compression and decompression algorithms for a given data type. In particular, it employs search techniques to mix and match different transformation, prediction, and encoding components to determine the most efficient configurations.