Big data analysis has become a crucial part of new emerging technologies such as the
internet of things, cyber-physical analysis, deep learning, anomaly detection, etc. Among …

Efficient numerical linear algebra is a core ingredient in many applications across almost all
scientific and industrial disciplines. With this survey we want to illustrate that numerical linear …

Post-training model compression can reduce the inference costs of deep neural networks,
but uncompressed training still consumes enormous hardware resources and energy. To …

Low Rank Decomposition (LRD) is a model compression technique applied to the weight
tensors of deep learning models in order to reduce the number of trainable parameters and …

Low rank decomposition decomposes each fully-connected layer of the transformer modules
into two smaller layers using Singular Value Decomposition. The state-of-the-art techniques …

We introduce a new method for speeding up the inference of deep neural networks. It is
somewhat inspired by the reduced-order modeling techniques for dynamical systems. The …

Summary Edge Artificial Intelligence (Edge AI) has become the buzzword for every industry
organization. Edge intelligence utilizes edge computing to access and analyze the data from …

Though deep neural networks achieve great accuracy in visual recognition tasks, they
contain millions of weights and thus require a large space to be stored. This presents a …

The rising computational and memory demands of machine learning models, particularly in
resource-constrained edge-device settings, motivate us to develop compressed models that …

Today neural networks are state of the art for numerous task including image classification.
The best neural networks use an enormous amount of parameters. Some works try to reduce …