The research activity is currently focused on algorithms, methodologies and tools for Internet of Things, Wearable Computing Systems, Edge AI, Cyber and IoT Security, Radiomics, Federated Learning, Blockchain, Intelligent Vehicles.
Internet of Things. Resp. Prof. Giancarlo Fortino
The Internet of Things (IoT) vision refers to a scenario in which daily real world objects participate to the Internet providing new cyber-physical services to both humans and machines, with consequent benefits for individuals and businesses. In such emerging scenario, Smart Objects (SOs), namely daily life physical object augmented with sensing/actuation, processing, storing, and networking capabilities, are fundamental building blocks to develop smart IoT applications and services atop multiple IoT platforms. The main research topics include the definition of methodologies and tools for analysis, design and implementation of dense and highly dynamic IoT ecosystems.
Edge Intelligence. Resp. Dr. Claudio Savaglio
The Edge Intelligence (EI) paradigm has recently emerged as a promising solution to overcome the inherent limitations of cloud computing (latency, autonomy, cost, data traffic, etc.) in the development and provision of next-generation Internet of Things (IoT) services. This truly distributed and pervasive computing approach brings artificial intelligence (AI) and machine learning (ML) technologies on edge devices, such as smartphones, IoT boards, cameras, for enabling real-time decision-making, improving scalability, privacy, and reliability, and ultimately, leading to more efficient and effective data analysis. The main addressed research topics in EI field pertains methodologies, architectures, models and simulators supporting the development of innovative EI systems spanning across the device-edge-cloud continuum, resorting to both on innovative approaches or on well-established ones, purposely customized for the EI domain.
Radiomics. Resp. Prof. Giancarlo Fortino
Radiomics is almost a one decade old scientific field, that uses quantitative analysis of medical images (such as CT, MR, and/or PET scans) in order to extract features, also called image biomarkers, which are able to reflect a clinical outcome. Developing concrete and effective radiomics applications is, however, a rather complex task. Indeed, the need of dealing with a wide range of heterogenous data sources, characterized by different kinds of dimensionality and features, variability of operators and the complexity of the analysis process which involves various radiomics tasks (es. Segmentation, feature extractions), pose several challenges. We are interested in developing novel and more accurate Machine/Deep Learning model for the several Radiomics tasks and, in particular in developing methods that can be used in clinical applications in many different ways, ranging from prognosis, diagnosis, prediction and/or analysis of the response to a certain therapy.
Federated Learning. Resp. Dr. Antonella Guzzo
Federated Learning is a novel collaborative and distributed learning paradigm, which enables training machine/deep learning models over remote devices or siloed data centers, such as mobile phones or hospitals, while keeping data localized. Differently from the distributed learning approach, federated learning provide a privacy- preserving solution by design, since the participant to the federation only share parameters /models without sharing data. Although the high potential of this approach, as in any emerging technologies, several challenges come into play. We are interested not only in developing solution that outperforms the existing ones, but also that address some specific challenges as dealing with Non-IID data, improving the level of sustainability and compliance with “green” issues.
Blockchain. Resp. Dr. Antonella Guzzo
Blockchain is a technology that provides user with desirable features of decentralization, autonomy, integrity, immutability, verification, fault-tolerance, anonymity, auditability, and transparency. In recent years, there was an explosion of initiative aiming at incorporate blockchain technology in several modern applications (in the healthcare, supply chain, finance etc.), and for each aforementioned sector a number of benefits, as well limits and challenges of the use of blockchain technologies emerge. We are interested in the development of new mechanisms and procedures of “accountability” based on DLT (Distributed Ledger Technology), blockchain and intelligent contracts (Smart Contract). Specifically, we focused on traceability of material goods through the use of Smart Tags and the definition of mechanisms for the recognition of the digital identity of physical objects for anti-countering.
Software Protection and IoT Security. Resp. Dr. Michele Ianni
In today’s digital era, software protection has become a critical necessity. As cyber-attacks become more advanced and widespread, it’s essential to consider obfuscation and other techniques for safeguarding software. Not only is it a priority for legitimate software developers to protect their intellectual property, but it is also significant for malware creators who strive to develop new viruses capable of evading antivirus detection. Attackers with malicious intentions also aim to steal sensitive information, compromise code integrity or insert backdoors into targeted systems. Binary analysis is an essential research topic that plays a crucial role in software protection and IoT security. By analyzing binary code to understand its functionality and behavior, binary analysis enables us to detect and eliminate vulnerabilities in software and hardware, which can be exploited by attackers. Furthermore, it can also help in identifying malicious code and protect systems against malware attacks.
Sustainable Cognitive Environments. Resp. Dr. Antonio Guerrieri
Sustainable Cognitive Environments (SCEs) represent a rapidly growing research area that aims to develop intelligent systems that are energy-efficient, environmentally friendly, and capable of adapting to the dynamically changing needs of users. By integrating the Internet of Things, Artificial Intelligence, and Edge Computing, SCEs can revolutionize how we interact with our surroundings and help us move towards a more sustainable future. In particular, Artificial Intelligence can enable these environments to learn from data, identify patterns, and make decisions without explicit human intervention. By leveraging advanced machine learning algorithms, SCEs can analyze large volumes of data from various sources, such as IoT devices or user inputs, to optimize energy usage, reduce waste, and improve overall sustainability. Moreover, SCE can also take advantage of Edge computing as it aims to process data closer to the source, reducing the need for constant communication with centralized data centers. By performing computation at the network’s edge, SCEs can significantly reduce energy consumption, latency, and bandwidth requirements, leading to more efficient and responsive systems. In conclusion, in the coming years, we can expect to see the widespread adoption of SCE paving the way toward a greener and more sustainable future for all.
Wearable Computing Systems. Resp. Dr. Raffaele Gravina
Thanks to the new wave of wearable gadgetry hitting the mass market, Wearable Computing Systems (WCS) are emerging as a new computing platform with full capabilities to support diversified application domains. Wearables, which include both devices and body sensors, are networked cyber-physical objects with enough power to support local computation of sensor-based information. To make the wearable devices connected and play a more significant role, the development of body sensor networks (BSNs) is essential. A BSN is a network of wireless wearable sensor nodes managed by more capable coordinators (smartphones, tablets, PCs). Although the basic elements (sensors, communication protocols, and software stacks) of a BSN are available, developing BSN systems/applications remains a complex task that requires design methods based on effective and efficient programming approaches and data processing techniques. BSNs involve wireless wearable physiological sensors applied to the human body for medical, wellness, and entertainment purposes. They allow for continuous, unobtrusive measurement of body movements and physiological signals, such as heart rate, muscular tension, skin conductivity, breathing rate, and volume, during the daily life of a user. The SPEME Lab has strong expertise and a long-history of interest in this domain, with significant results obtained under the umbrella of the SPINE Body-of-Knowledge (https://spine-bok.dimes.unical.it) in the context of BSN/WCS design methods, architectures, software prototyping tools, processing algorithms and machine learning-based analysis of physiological signals.