Research into optimisation for deep learning is characterised by a tension between the
computational efficiency of first-order, gradient-based methods (such as SGD and Adam) …

As data sets grow in size and complexity, it is becoming more difficult to pull useful features
from them using hand-crafted feature extractors. For this reason, deep learning (DL) …

Reliability optimization can be applied to find parameters that increase reliability and
decrease costs, in the presence of uncertainty. Nowadays, with the increasing complexity of …

In an age defined by explosive growth in information technology, data generation, storage
and transmission have increased dramatically. This data fuels the core of machine learning …

This work proposes a novel application of momentum to the L-BFGS Quasi-Newton method
and seeks to evaluate its performance when applied to optimization problems. We …

Distributed machine learning tasks often involve complex device interactions that can lead to
high energy consumption. This work investigates optimal device placement strategies to …

Stochastic Gradient Descent is an algorithm for optimization. Compared to Batch methods,
there are some motivations for using stochastic methods. It is practical (when the dataset is …

Abstract The YOLOv5 (You Only Look Once version 5) network has gained attention for its
performance in object detection tasks. This study focuses on optimizing YOLOv5 across …

Distributed networks face significant challenges related to latency and bandwidth when it
comes to real-time data sharing and processing. To tackle these issues, we present the Real …