TY - JOUR AU - Mokadem, Frank AU - Wong, Alexander PY - 2023/05/10 Y2 - 2024/03/29 TI - Automated search for optimal convolutional neural network factorization JF - Journal of Computational Vision and Imaging Systems JA - J. Comp. Vis. Imag. Sys. VL - 8 IS - 1 SE - Articles DO - 10.15353/jcvis.v8i1.5379 UR - https://openjournals.uwaterloo.ca/index.php/vsl/article/view/5379 SP - 62-63 AB - <table style="height: 215px;" width="537"><tbody><tr><td width="711">Deep Neural networks (DNNs) are the state of the art technique<br>when it comes to artificial intelligence tasks relating to computer vision. Usage of DNNs is wide spread across multiple industries, and<br>entertainment. Most notably is the use of Convolutional Neural Networks(CNNs) architectures for object detection and classification,<br>and even more recently information retrieval. However, one downfall of CNNs is their computational cost, even on trivial tasks. The<br>reason for such high computational cost lies in the high flop number of kernel convolution, the core operation which is built upon a<br>CNN. Hence presenting the need for compression of floating number information in CNNs. In this work we explore the techniques of<br>CNN compression using tensor decompositions. Furtheremore, we<br>aspire to build an automated tool to execute a grid search through<br>the space of all possible factorizations of a CNN and pick an optimal<br>compressed network representation with respect to performance requirements.</td></tr></tbody></table> ER -