Deep learning models for the automatic analysis of volleyball plays

Collecting and analysing data is an important element in any field of human activity and research. Even in sports, collecting and analyzing statistical data is attracting a growing interest. Some exemplar use cases are: improvement of technical/tactical aspects for team coaches, definition of game strategies based on the opposite team play or evaluation of the performance of players. Other advantages are related to taking more precise and impartial judgment in referee decisions: a wrong decision can change the outcomes of important matches. Finally, it can be useful to provide better representations and graphic effects that make the game more engaging for the audience during the match. Nowadays it is possible to delegate this type of task to automatic software systems that can use cameras or even hardware sensors to collect images or data and process them. One of the most efficient methods to collect data is to process the video images of the sporting event through mixed techniques concerning machine learning applied to computer vision. As in other domains in which computer vision can be applied, the main tasks in sports are related to object detection, player tracking, and to the pose estimation of athletes. The goal of the present thesis is to apply different models of CNNs to analyze volleyball matches. Starting from video frames of a volleyball match, we reproduce a bird's eye view of the playing court where all the players are projected, reporting also for each player the type of action she/he is performing.

Model-Based validation of Driver Drowsiness Detection System for ADAS

The work described in this Master’s Degree thesis was born after the collaboration with the company Maserati S.p.a, an Italian luxury car maker with its headquarters located in Modena, in the heart of the Italian Motor Valley, where I worked as a stagiaire in the Virtual Engineering team between September 2021 and February 2022. This work proposes the validation using real-world ECUs of a Driver Drowsiness Detection (DDD) system prototype based on different detection methods with the goal to overcome input signal losses and system failures. Detection methods of different categories have been chosen from literature and merged with the goal of utilizing the benefits of each of them, overcoming their limitations and limiting as much as possible their degree of intrusiveness to prevent any kind of driving distraction: an image processing-based technique for human physical signals detection as well as methods based on driver-vehicle interaction are used. A Driver-In-the-Loop simulator is used to gather real data on which a Machine Learning-based algorithm will be trained and validated. These data come from the tests that the company conducts in its daily activities so confidential information about the simulator and the drivers will be omitted. Although the impact of the proposed system is not remarkable and there is still work to do in all its elements, the results indicate the main advantages of the system in terms of robustness against subsystem failures and signal losses.

End-to-end Deep Metric Learning con Vision-Language Model per il Fashion Image Captioning

L'image captioning è un task di machine learning che consiste nella generazione di una didascalia, o caption, che descriva le caratteristiche di un'immagine data in input. Questo può essere applicato, ad esempio, per descrivere in dettaglio i prodotti in vendita su un sito di e-commerce, migliorando l'accessibilità del sito web e permettendo un acquisto più consapevole ai clienti con difficoltà visive. La generazione di descrizioni accurate per gli articoli di moda online è importante non solo per migliorare le esperienze di acquisto dei clienti, ma anche per aumentare le vendite online. Oltre alla necessità di presentare correttamente gli attributi degli articoli, infatti, descrivere i propri prodotti con il giusto linguaggio può contribuire a catturare l'attenzione dei clienti. In questa tesi, ci poniamo l'obiettivo di sviluppare un sistema in grado di generare una caption che descriva in modo dettagliato l'immagine di un prodotto dell'industria della moda dato in input, sia esso un capo di vestiario o un qualche tipo di accessorio. A questo proposito, negli ultimi anni molti studi hanno proposto soluzioni basate su reti convoluzionali e LSTM. In questo progetto proponiamo invece un'architettura encoder-decoder, che utilizza il modello Vision Transformer per la codifica delle immagini e GPT-2 per la generazione dei testi. Studiamo inoltre come tecniche di deep metric learning applicate in end-to-end durante l'addestramento influenzino le metriche e la qualità delle caption generate dal nostro modello.

Design and implementation of a Reinforcement Learning framework for iOS devices

Reinforcement Learning is an increasingly popular area of Artificial Intelligence. The applications of this learning paradigm are many, but its application in mobile computing is in its infancy. This study aims to provide an overview of current Reinforcement Learning applications on mobile devices, as well as to introduce a new framework for iOS devices: Swift-RL Lib. This new Swift package allows developers to easily support and integrate two of the most common RL algorithms, Q-Learning and Deep Q-Network, in a fully customizable environment. All processes are performed on the device, without any need for remote computation. The framework was tested in different settings and evaluated through several use cases. Through an in-depth performance analysis, we show that the platform provides effective and efficient support for Reinforcement Learning for mobile applications.

Edge-aware Graph Attention Networks: Joint Reasoning On Text and Knowledge Graphs for Biomedical Question Answering

Much of the real-world dataset, including textual data, can be represented using graph structures. The use of graphs to represent textual data has many advantages, mainly related to maintaining a more significant amount of information, such as the relationships between words and their types. In recent years, many neural network architectures have been proposed to deal with tasks on graphs. Many of them consider only node features, ignoring or not giving the proper relevance to relationships between them. However, in many node classification tasks, they play a fundamental role. This thesis aims to analyze the main GNNs, evaluate their advantages and disadvantages, propose an innovative solution considered as an extension of GAT, and apply them to a case study in the biomedical field. We propose the reference GNNs, implemented with methodologies later analyzed, and then applied to a question answering system in the biomedical field as a replacement for the pre-existing GNN. We attempt to obtain better results by using models that can accept as input both node and edge features. As shown later, our proposed models can beat the original solution and define the state-of-the-art for the task under analysis.

Hybrid Artificial Intelligence to Extract Patterns and Rules from Argumentative and Legal Texts

This Thesis is composed of a collection of works written in the period 2019-2022, whose aim is to find methodologies of Artificial Intelligence (AI) and Machine Learning to detect and classify patterns and rules in argumentative and legal texts. We define our approach “hybrid”, since we aimed at designing hybrid combinations of symbolic and sub-symbolic AI, involving both “top-down” structured knowledge and “bottom-up” data-driven knowledge. A first group of works is dedicated to the classification of argumentative patterns. Following the Waltonian model of argument and the related theory of Argumentation Schemes, these works focused on the detection of argumentative support and opposition, showing that argumentative evidences can be classified at fine-grained levels without resorting to highly engineered features. To show this, our methods involved not only traditional approaches such as TFIDF, but also some novel methods based on Tree Kernel algorithms. After the encouraging results of this first phase, we explored the use of a some emerging methodologies promoted by actors like Google, which have deeply changed NLP since 2018-19 — i.e., Transfer Learning and language models. These new methodologies markedly improved our previous results, providing us with best-performing NLP tools. Using Transfer Learning, we also performed a Sequence Labelling task to recognize the exact span of argumentative components (i.e., claims and premises), thus connecting portions of natural language to portions of arguments (i.e., to the logical-inferential dimension). The last part of our work was finally dedicated to the employment of Transfer Learning methods for the detection of rules and deontic modalities. In this case, we explored a hybrid approach which combines structured knowledge coming from two LegalXML formats (i.e., Akoma Ntoso and LegalRuleML) with sub-symbolic knowledge coming from pre-trained (and then fine-tuned) neural architectures.

Embedded representations of social interactions

Social interactions have been the focus of social science research for a century, but their study has recently been revolutionized by novel data sources and by methods from computer science, network science, and complex systems science. The study of social interactions is crucial for understanding complex societal behaviours. Social interactions are naturally represented as networks, which have emerged as a unifying mathematical language to understand structural and dynamical aspects of socio-technical systems. Networks are, however, highly dimensional objects, especially when considering the scales of real-world systems and the need to model the temporal dimension. Hence the study of empirical data from social systems is challenging both from a conceptual and a computational standpoint. A possible approach to tackling such a challenge is to use dimensionality reduction techniques that represent network entities in a low-dimensional feature space, preserving some desired properties of the original data. Low-dimensional vector space representations, also known as network embeddings, have been extensively studied, also as a way to feed network data to machine learning algorithms. Network embeddings were initially developed for static networks and then extended to incorporate temporal network data. We focus on dimensionality reduction techniques for time-resolved social interaction data modelled as temporal networks. We introduce a novel embedding technique that models the temporal and structural similarities of events rather than nodes. Using empirical data on social interactions, we show that this representation captures information relevant for the study of dynamical processes unfolding over the network, such as epidemic spreading. We then turn to another large-scale dataset on social interactions: a popular Web-based crowdfunding platform. We show that tensor-based representations of the data and dimensionality reduction techniques such as tensor factorization allow us to uncover the structural and temporal aspects of the system and to relate them to geographic and temporal activity patterns.

Big data analytics for proactive and predictive maintenance in electric car battery packs

The idea behind the project is to develop a methodology for analyzing and developing techniques for the diagnosis and the prediction of the state of charge and health of lithium-ion batteries for automotive applications. For lithium-ion batteries, residual functionality is measured in terms of state of health; however, this value cannot be directly associated with a measurable value, so it must be estimated. The development of the algorithms is based on the identification of the causes of battery degradation, in order to model and predict the trend. Therefore, models have been developed that are able to predict the electrical, thermal and aging behavior. In addition to the model, it was necessary to develop algorithms capable of monitoring the state of the battery, online and offline. This was possible with the use of algorithms based on Kalman filters, which allow the estimation of the system status in real time. Through machine learning algorithms, which allow offline analysis of battery deterioration using a statistical approach, it is possible to analyze information from the entire fleet of vehicles. Both systems work in synergy in order to achieve the best performance. Validation was performed with laboratory tests on different batteries and under different conditions. The development of the model allowed to reduce the time of the experimental tests. Some specific phenomena were tested in the laboratory, and the other cases were artificially generated.

Exploiting the environment for structural health monitoring: Strategies, algorithms, and technologies

A densely built environment is a complex system of infrastructure, nature, and people closely interconnected and interacting. Vehicles, public transport, weather action, and sports activities constitute a manifold set of excitation and degradation sources for civil structures. In this context, operators should consider different factors in a holistic approach for assessing the structural health state. Vibration-based structural health monitoring (SHM) has demonstrated great potential as a decision-supporting tool to schedule maintenance interventions. However, most excitation sources are considered an issue for practical SHM applications since traditional methods are typically based on strict assumptions on input stationarity. Last-generation low-cost sensors present limitations related to a modest sensitivity and high noise floor compared to traditional instrumentation. If these devices are used for SHM in urban scenarios, short vibration recordings collected during high-intensity events and vehicle passage may be the only available datasets with a sufficient signal-to-noise ratio. While researchers have spent efforts to mitigate the effects of short-term phenomena in vibration-based SHM, the ultimate goal of this thesis is to exploit them and obtain valuable information on the structural health state. First, this thesis proposes strategies and algorithms for smart sensors operating individually or in a distributed computing framework to identify damage-sensitive features based on instantaneous modal parameters and influence lines. Ordinary traffic and people activities become essential sources of excitation, while human-powered vehicles, instrumented with smartphones, take the role of roving sensors in crowdsourced monitoring strategies. The technical and computational apparatus is optimized using in-memory computing technologies. Moreover, identifying additional local features can be particularly useful to support the damage assessment of complex structures. Thereby, smart coatings are studied to enable the self-sensing properties of ordinary structural elements. In this context, a machine-learning-aided tomography method is proposed to interpret the data provided by a nanocomposite paint interrogated electrically.

Big Code Applications and Approaches

The availability of a huge amount of source code from code archives and open-source projects opens up the possibility to merge machine learning, programming languages, and software engineering research fields. This area is often referred to as Big Code where programming languages are treated instead of natural languages while different features and patterns of code can be exploited to perform many useful tasks and build supportive tools. Among all the possible applications which can be developed within the area of Big Code, the work presented in this research thesis mainly focuses on two particular tasks: the Programming Language Identification (PLI) and the Software Defect Prediction (SDP) for source codes. Programming language identification is commonly needed in program comprehension and it is usually performed directly by developers. However, when it comes at big scales, such as in widely used archives (GitHub, Software Heritage), automation of this task is desirable. To accomplish this aim, the problem is analyzed from different points of view (text and image-based learning approaches) and different models are created paying particular attention to their scalability. Software defect prediction is a fundamental step in software development for improving quality and assuring the reliability of software products. In the past, defects were searched by manual inspection or using automatic static and dynamic analyzers. Now, the automation of this task can be tackled using learning approaches that can speed up and improve related procedures. Here, two models have been built and analyzed to detect some of the commonest bugs and errors at different code granularity levels (file and method levels). Exploited data and models’ architectures are analyzed and described in detail. Quantitative and qualitative results are reported for both PLI and SDP tasks while differences and similarities concerning other related works are discussed.

Continual learning for computer vision applications

One of the most visionary goals of Artificial Intelligence is to create a system able to mimic and eventually surpass the intelligence observed in biological systems including, ambitiously, the one observed in humans. The main distinctive strength of humans is their ability to build a deep understanding of the world by learning continuously and drawing from their experiences. This ability, which is found in various degrees in all intelligent biological beings, allows them to adapt and properly react to changes by incrementally expanding and refining their knowledge. Arguably, achieving this ability is one of the main goals of Artificial Intelligence and a cornerstone towards the creation of intelligent artificial agents. Modern Deep Learning approaches allowed researchers and industries to achieve great advancements towards the resolution of many long-standing problems in areas like Computer Vision and Natural Language Processing. However, while this current age of renewed interest in AI allowed for the creation of extremely useful applications, a concerningly limited effort is being directed towards the design of systems able to learn continuously. The biggest problem that hinders an AI system from learning incrementally is the catastrophic forgetting phenomenon. This phenomenon, which was discovered in the 90s, naturally occurs in Deep Learning architectures where classic learning paradigms are applied when learning incrementally from a stream of experiences. This dissertation revolves around the Continual Learning field, a sub-field of Machine Learning research that has recently made a comeback following the renewed interest in Deep Learning approaches. This work will focus on a comprehensive view of continual learning by considering algorithmic, benchmarking, and applicative aspects of this field. This dissertation will also touch on community aspects such as the design and creation of research tools aimed at supporting Continual Learning research, and the theoretical and practical aspects concerning public competitions in this field.

Concepts and methods for efficient decentralized learning in federated settings

Deep Neural Networks (DNNs) have revolutionized a wide range of applications beyond traditional machine learning and artificial intelligence fields, e.g., computer vision, healthcare, natural language processing and others. At the same time, edge devices have become central in our society, generating an unprecedented amount of data which could be used to train data-hungry models such as DNNs. However, the potentially sensitive or confidential nature of gathered data poses privacy concerns when storing and processing them in centralized locations. To this purpose, decentralized learning decouples model training from the need of directly accessing raw data, by alternating on-device training and periodic communications. The ability of distilling knowledge from decentralized data, however, comes at the cost of facing more challenging learning settings, such as coping with heterogeneous hardware and network connectivity, statistical diversity of data, and ensuring verifiable privacy guarantees. This Thesis proposes an extensive overview of decentralized learning literature, including a novel taxonomy and a detailed description of the most relevant system-level contributions in the related literature for privacy, communication efficiency, data and system heterogeneity, and poisoning defense. Next, this Thesis presents the design of an original solution to tackle communication efficiency and system heterogeneity, and empirically evaluates it on federated settings. For communication efficiency, an original method, specifically designed for Convolutional Neural Networks, is also described and evaluated against the state-of-the-art. Furthermore, this Thesis provides an in-depth review of recently proposed methods to tackle the performance degradation introduced by data heterogeneity, followed by empirical evaluations on challenging data distributions, highlighting strengths and possible weaknesses of the considered solutions. Finally, this Thesis presents a novel perspective on the usage of Knowledge Distillation as a mean for optimizing decentralized learning systems in settings characterized by data heterogeneity or system heterogeneity. Our vision on relevant future research directions close the manuscript.

Development of unsupervised learning methods with applications to life sciences data

Machine Learning makes computers capable of performing tasks typically requiring human intelligence. A domain where it is having a considerable impact is the life sciences, allowing to devise new biological analysis protocols, develop patients’ treatments efficiently and faster, and reduce healthcare costs. This Thesis work presents new Machine Learning methods and pipelines for the life sciences focusing on the unsupervised field. At a methodological level, two methods are presented. The first is an “Ab Initio Local Principal Path” and it is a revised and improved version of a pre-existing algorithm in the manifold learning realm. The second contribution is an improvement over the Import Vector Domain Description (one-class learning) through the Kullback-Leibler divergence. It hybridizes kernel methods to Deep Learning obtaining a scalable solution, an improved probabilistic model, and state-of-the-art performances. Both methods are tested through several experiments, with a central focus on their relevance in life sciences. Results show that they improve the performances achieved by their previous versions. At the applicative level, two pipelines are presented. The first one is for the analysis of RNA-Seq datasets, both transcriptomic and single-cell data, and is aimed at identifying genes that may be involved in biological processes (e.g., the transition of tissues from normal to cancer). In this project, an R package is released on CRAN to make the pipeline accessible to the bioinformatic Community through high-level APIs. The second pipeline is in the drug discovery domain and is useful for identifying druggable pockets, namely regions of a protein with a high probability of accepting a small molecule (a drug). Both these pipelines achieve remarkable results. Lastly, a detour application is developed to identify the strengths/limitations of the “Principal Path” algorithm by analyzing Convolutional Neural Networks induced vector spaces. This application is conducted in the music and visual arts domains.

Novel methods of state estimation and parameter identification for high-performance motorcycles

This research activity aims at providing a reliable estimation of particular state variables or parameters concerning the dynamics and performance optimization of a MotoGP-class motorcycle, integrating the classical model-based approach with new methodologies involving artificial intelligence. The first topic of the research focuses on the estimation of the thermal behavior of the MotoGP carbon braking system. Numerical tools are developed to assess the instantaneous surface temperature distribution in the motorcycle's front brake discs. Within this application other important brake parameters are identified using Kalman filters, such as the disc convection coefficient and the power distribution in the disc-pads contact region. Subsequently, a physical model of the brake is built to estimate the instantaneous braking torque. However, the results obtained with this approach are highly limited by the knowledge of the friction coefficient (μ) between the disc rotor and the pads. Since the value of μ is a highly nonlinear function of many variables (namely temperature, pressure and angular velocity of the disc), an analytical model for the friction coefficient estimation appears impractical to establish. To overcome this challenge, an innovative hybrid solution is implemented, combining the benefit of artificial intelligence (AI) with classical model-based approach. Indeed, the disc temperature estimated through the thermal model previously implemented is processed by a machine learning algorithm that outputs the actual value of the friction coefficient thus improving the braking torque computation performed by the physical model of the brake. Finally, the last topic of this research activity regards the development of an AI algorithm to estimate the current sideslip angle of the motorcycle's front tire. While a single-track motorcycle kinematic model and IMU accelerometer signals theoretically enable sideslip calculation, the presence of accelerometer noise leads to a significant drift over time. To address this issue, a long short-term memory (LSTM) network is implemented.

Identificazione e classificazione delle lesioni endocavitarie uterine dalle immagini isteroscopiche con il deep learning

Anche se l'isteroscopia con la biopsia endometriale è il gold standard nella diagnosi della patologia intracavitaria uterina, l'esperienza dell’isteroscopista è fondamentale per una diagnosi corretta. Il Deep Learning (DL) come metodica di intelligenza artificiale potrebbe essere un aiuto per superare questo limite. Sono disponibili pochi studi con risultati preliminari e mancano ricerche che valutano le prestazioni dei modelli di DL nell'identificazione delle lesioni intrauterine e il possibile aiuto derivato dai fattori clinici. Obiettivo: Sviluppare un modello di DL per identificare e classificare le patologie endocavitarie uterine dalle immagini isteroscopiche. Metodi: È stato eseguito uno studio di coorte retrospettivo osservazionale monocentrico su una serie consecutiva di casi isteroscopici di pazienti con patologia intracavitaria uterina confermata all’esame istologico eseguiti al Policlinico S. Orsola. Le immagini isteroscopiche sono state usate per costruire un modello di DL per la classificazione e l'identificazione delle lesioni intracavitarie con e senza l'aiuto di fattori clinici (età, menopausa, AUB, terapia ormonale e tamoxifene). Come risultati dello studio abbiamo calcolato le metriche diagnostiche del modello di DL nella classificazione e identificazione delle lesioni uterine intracavitarie con e senza l'aiuto dei fattori clinici. Risultati: Abbiamo esaminato 1.500 immagini provenienti da 266 casi: 186 pazienti avevano lesioni focali benigne, 25 lesioni diffuse benigne e 55 lesioni preneoplastiche/neoplastiche. Sia per quanto riguarda la classificazione che l’identificazione, le migliori prestazioni sono state raggiunte con l'aiuto dei fattori clinici, complessivamente con precision dell'80,11%, recall dell'80,11%, specificità del 90,06%, F1 score dell’80,11% e accuratezza dell’86,74% per la classificazione. Per l’identificazione abbiamo ottenuto un rilevamento complessivo dell’85,82%, precision 93,12%, recall del 91,63% ed F1 score del 92,37%. Conclusioni: Il modello DL ha ottenuto una bassa performance nell’identificazione e classificazione delle lesioni intracavitarie uterine dalle immagini isteroscopiche. Anche se la migliore performance diagnostica è stata ottenuta con l’aiuto di fattori clinici specifici, questo miglioramento è stato scarso.