Publication Details
Genetic Programming with Memory for Approximate Data Reconstruction
Jůza Tadeáš, Ing.
genetic programming, convolutional neural network, approximate computing,
hardware accelerator, classification, energy
This chapter addresses the computation-memorization trade-offs in the context of
genetic programming (GP). We introduce genetic programming with memory (GPM) in
which GP evolves not only the expression but also the content of a small local
memory to better approximate the original data set. In particular, we evolved
expression-memory pairs that can serve as weight generators and thus approximate
the weights associated with convolutional layers of some convolutional neural
networks (CNNs). This is potentially interesting for the efficient
implementations of hardware accelerators of CNNs in which memory access is
significantly more energy-demanding than arithmetic operations. In our approach,
most of the weights are approximated using an evolved expression; only some
fraction of them must be read from memory. For example, if memory contains 10% of
the original weights, the weight generator evolved for a convolutional layer can
approximate the original weights such that the CNN utilizing the generated
weights shows less than a 1% drop in the classification accuracy on the MNIST
data set. The memory requirements are reduced 3.1x or 12.6x for 8-bit or 32-bit
weights, respectively. Additional experiments conducted for more complex CNNs and
challenging image classification benchmarks show various impacts of weights'
approximation on classification accuracy.
@inbook{BUT193318,
author="Lukáš {Sekanina} and Tadeáš {Jůza}",
title="Genetic Programming with Memory for Approximate Data Reconstruction",
booktitle="Genetic Programming Theory and Practice XXI",
year="2025",
publisher="Springer Nature Singapore",
address="Singapore",
pages="199--218",
doi="10.1007/978-981-96-0077-9\{_}10",
isbn="978-981-9600-76-2",
url="https://link.springer.com/chapter/10.1007/978-981-96-0077-9_10"
}