arXiv subjects

Lorenzo Valerio

Publications and source records attributed to Lorenzo Valerio.

26 records · Page 2Linked to original sources

arXiv2022

Federated Feature Selection for Cyber-Physical Systems of Systems

Autonomous vehicles (AVs) generate a massive amount of multi-modal data that once collected and processed through Machine Learning algorithms, enable AI-based services at the Edge. In fact, not all these data contain valuable, and informative content but only a subset of the relative attributes should be exploited at the Edge. Therefore, enabling AVs to locally extract such a subset is of utmost importance to limit computation and communication workloads. Achieving a consistent subset of data in a distributed manner imposes the AVs to cooperate in finding an agreement on what attributes should be sent to the Edge. In this work, we address such a problem by proposing a federated feature selection algorithm where all the AVs collaborate to filter out, iteratively, the redundant or irrelevant attributes in a distributed manner, without any exchange of raw data. This solution builds on two components: a Mutual-Information-based feature selection algorithm run by the AVs and a novel aggregation function based on the Bayes theorem executed on the Edge. Our federated feature selection algorithm provably converges to a solution in a finite number of steps. Such an algorithm has been tested on two reference datasets: MAV with images and inertial measurements of a monitored vehicle, WESAD with a collection of samples from biophysical sensors to monitor a relative passenger. The numerical results show that the fleet finds a consensus with both the datasets on the minimum achievable subset of features, i.e., 24 out of 2166 (99\%) in MAV and 4 out of 8 (50\%) in WESAD, preserving the informative content of data.

cs.LG↗

arXiv2022

Federated Semi-Supervised Classification of Multimedia Flows for 3D Networks

Automatic traffic classification is increasingly becoming important in traffic engineering, as the current trend of encrypting transport information (e.g., behind HTTP-encrypted tunnels) prevents intermediate nodes from accessing end-to-end packet headers. However, this information is crucial for traffic shaping, network slicing, and Quality of Service (QoS) management, for preventing network intrusion, and for anomaly detection. 3D networks offer multiple routes that can guarantee different levels of QoS. Therefore, service classification and separation are essential to guarantee the required QoS level to each traffic sub-flow through the appropriate network trunk. In this paper, a federated feature selection and feature reduction learning scheme is proposed to classify network traffic in a semi-supervised cooperative manner. The federated gateways of 3D network help to enhance the global knowledge of network traffic to improve the accuracy of anomaly and intrusion detection and service identification of a new traffic flow.

cs.LG↗

arXiv2021

Dynamic Hard Pruning of Neural Networks at the Edge of the Internet

Neural Networks (NN), although successfully applied to several Artificial Intelligence tasks, are often unnecessarily over-parametrised. In edge/fog computing, this might make their training prohibitive on resource-constrained devices, contrasting with the current trend of decentralising intelligence from remote data centres to local constrained devices. Therefore, we investigate the problem of training effective NN models on constrained devices having a fixed, potentially small, memory budget. We target techniques that are both resource-efficient and performance effective while enabling significant network compression. Our Dynamic Hard Pruning (DynHP) technique incrementally prunes the network during training, identifying neurons that marginally contribute to the model accuracy. DynHP enables a tunable size reduction of the final neural network and reduces the NN memory occupancy during training. Freed memory is reused by a \emph{dynamic batch sizing} approach to counterbalance the accuracy degradation caused by the hard pruning strategy, improving its convergence and effectiveness. We assess the performance of DynHP through reproducible experiments on three public datasets, comparing them against reference competitors. Results show that DynHP compresses a NN up to $10$ times without significant performance drops (up to $3.5\%$ additional error w.r.t. the competitors), reducing up to $80\%$ the training memory occupancy.

cs.LG↗

arXiv2021

Cellular traffic offloading via Opportunistic Networking with Reinforcement Learning

The widespread diffusion of mobile phones is triggering an exponential growth of mobile data traffic that is likely to cause, in the near future, considerable traffic overload issues even in last-generation cellular networks. Offloading part of the traffic to other networks is considered a very promising approach and, in particular, in this paper, we consider offloading through opportunistic networks of users' devices. However, the performance of this solution strongly depends on the pattern of encounters between mobile nodes, which should therefore be taken into account when designing offloading control algorithms. In this paper, we propose an adaptive offloading solution based on the Reinforcement Learning framework and we evaluate and compare the performance of two well-known learning algorithms: Actor-Critic and Q-Learning. More precisely, in our solution the controller of the dissemination process, once trained, is able to select a proper number of content replicas to be injected into the opportunistic network to guarantee the timely delivery of contents to all interested users. We show that our system based on Reinforcement Learning is able to automatically learn a very efficient strategy to reduce the traffic on the cellular network, without relying on any additional context information about the opportunistic network. Our solution achieves a higher level of offloading with respect to other state-of-the-art approaches, in a range of different mobility settings. Moreover, we show that a more refined learning solution, based on the Actor-Critic algorithm, is significantly more efficient than a simpler solution based on Q-learning.

cs.NI↗

arXiv2021

A communication efficient distributed learning framework for smart environments

Due to the pervasive diffusion of personal mobile and IoT devices, many ``smart environments'' (e.g., smart cities and smart factories) will be, among others, generators of huge amounts of data. Currently, this is typically achieved through centralised cloud-based data analytics services. However, according to many studies, this approach may present significant issues from the standpoint of data ownership, and even wireless network capacity. One possibility to cope with these shortcomings is to move data analytics closer to where data is generated. In this paper, we tackle this issue by proposing and analyzing a distributed learning framework, whereby data analytics are performed at the edge of the network, i.e., on locations very close to where data is generated. Specifically, in our framework, partial data analytics are performed directly on the nodes that generate the data, or on nodes close by (e.g., some of the data generators can take this role on behalf of subsets of other nodes nearby). Then, nodes exchange partial models and refine them accordingly. Our framework is general enough to host different analytics services. In the specific case analysed in the paper, we focus on a learning task, considering two distributed learning algorithms. Using an activity recognition and a pattern recognition task, both on reference datasets, we compare the two learning algorithms between each other and with a central cloud solution (i.e., one that has access to the complete datasets). Our results show that using distributed machine learning techniques, it is possible to drastically reduce the network overhead, while obtaining performance comparable to the cloud solution in terms of learning accuracy. The analysis also shows when each distributed learning approach is preferable, based on the specific distribution of the data on the nodes.

cs.DC↗

arXiv2021

Energy efficient distributed analytics at the edge of the network for IoT environments

Due to the pervasive diffusion of personal mobile and IoT devices, many "smart environments" (e.g., smart cities and smart factories) will be, generators of huge amounts of data. Currently, analysis of this data is typically achieved through centralised cloud-based services. However, according to many studies, this approach may present significant issues from the standpoint of data ownership, as well as wireless network capacity. In this paper, we exploit the fog computing paradigm to move computation close to where data is produced. We exploit a well-known distributed machine learning framework (Hypothesis Transfer Learning), and perform data analytics on mobile nodes passing by IoT devices, in addition to fog gateways at the edge of the network infrastructure. We analyse the performance of different configurations of the distributed learning framework, in terms of (i) accuracy obtained in the learning task and (ii) energy spent to send data between the involved nodes. Specifically, we consider reference wireless technologies for communication between the different types of nodes we consider, e.g. LTE, Nb-IoT, 802.15.4, 802.11, etc. Our results show that collecting data through the mobile nodes and executing the distributed analytics using short-range communication technologies, such as 802.15.4 and 802.11, allows to strongly reduce the energy consumption of the system up to $94\%$ with a loss in accuracy w.r.t. a centralised cloud solution up to $2\%$.

cs.DC↗

arXiv2021

Optimising cost vs accuracy of decentralised analytics in fog computing environments

The exponential growth of devices and data at the edges of the Internet is rising scalability and privacy concerns on approaches based exclusively on remote cloud platforms. Data gravity, a fundamental concept in Fog Computing, points towards decentralisation of computation for data analysis, as a viable alternative to address those concerns. Decentralising AI tasks on several cooperative devices means identifying the optimal set of locations or Collection Points (CP for short) to use, in the continuum between full centralisation (i.e., all data on a single device) and full decentralisation (i.e., data on source locations). We propose an analytical framework able to find the optimal operating point in this continuum, linking the accuracy of the learning task with the corresponding network and computational cost for moving data and running the distributed training at the CPs. We show through simulations that the model accurately predicts the optimal trade-off, quite often an intermediate point between full centralisation and full decentralisation, showing also a significant cost saving w.r.t. both of them. Finally, the analytical model admits closed-form or numeric solutions, making it not only a performance evaluation instrument but also a design tool to configure a given distributed learning task optimally before its deployment.

cs.DC↗

arXiv2013

Application of a cognitive-inspired algorithm for detecting communities in mobility networks

The emergence and the global adaptation of mobile devices has influenced human interactions at the individual, community, and social levels leading to the so called Cyber-Physical World (CPW) convergence scenario [1]. One of the most important features of CPW is the possibility of exploiting information about the structure of the social communities of users, revealed by joint movement patterns and frequency of physical co-location. Mobile devices of users that belong to the same social community are likely to "see" each other (and thus be able to communicate through ad-hoc networking techniques) more frequently and regularly than devices outside the community. In mobile opportunistic networks, this fact can be exploited, for example, to optimize networking operations such as forwarding and dissemination of messages. In this paper we present the application of a cognitive-inspired algorithm [2,3,4] for revealing the structure of these dynamic social networks (simulated by the HCMM model [5]) using information about physical encounters logged by the users' mobile devices. The main features of our algorithm are: (i) the capacity of detecting social communities induced by physical co-location of users through distributed algorithms; (ii) the capacity to detect users belonging to more communities (thus acting as bridges across them), and (iii) the capacity to detect the time evolution of communities.

physics.soc-ph↗