Internet of Things and Deep Learning are synergetically and exponentially growing industrial fields with a massive call for their unification into a common framework called Edge AI. While on-device inference is a well-explored topic in recent research, backpropagation remains an open challenge due to its prohibitive computational and memory costs compared to the extreme resource constraints of embedded devices. Drawing on tensor decomposition research, we tackle the main bottleneck of backpropagation, namely the memory footprint of activation map storage. We investigate and compare the effects of activation compression using Singular Value Decomposition and its tensor variant, High-Order Singular Value Decomposition. The application of low-order decomposition results in considerable memory savings while preserving the features essential for learning, and also offers theoretical guarantees to convergence. Experimental results obtained on main-stream architectures and tasks demonstrate Pareto-superiority over other state-of-the-art solutions, in terms of the trade-off between generalization and memory footprint.

In this paper we examine how to efficiently build a brokers overlay to implement publish/subscribe in a wireless sensor network, trying to reduce sensor nodes energy consumption, memory required for buffering packets and delivery time. We evaluate, on the connectivity graph that represents the network, the performance of various criteria that can be used to select brokers among the set of nodes. We compare a dominating set approach with the selection of the best ranked brokers based on centrality measures. We finally give hints on how to implement a distributed algorithm to approximate the most efficient overlays.

This paper describes how GLADE, our implementation of the Ada 95 Distributed Systems Annex, can be used to build large object-oriented real-time distributed systems. In addition to the powerful distribution features included in the Ada 95 language itself, we provide extensions to help the programmer build robust and failsafe distributed applications.