The development of selective copying: children’s learning from an expert versus their mother

This study tested the prediction that, with age, children should rely less on familiarity and more on expertise in their selective social learning. Experiment 1 (N=50) found that 5- to 6-year-olds copied the technique their mother used to extract a prize from a novel puzzle box, in preference to both a stranger and an established expert. This bias occurred despite children acknowledging the expert model’s superior capability. Experiment 2 (N=50) demonstrated a shift in 7-to 8-year-olds towards copying the expert. Children aged 9- to 10-years did not copy according to a model bias. The findings of a follow-up study (N=30) confirmed that, instead, they prioritized their own – partially flawed – causal understanding of the puzzle box.

A machine learning approach for recognising woody plants on railway trackbeds

The purpose of this work in progress study was to test the concept of recognising plants using images acquired by image sensors in a controlled noise-free environment. The presence of vegetation on railway trackbeds and embankments presents potential problems. Woody plants (e.g. Scots pine, Norway spruce and birch) often establish themselves on railway trackbeds. This may cause problems because legal herbicides are not effective in controlling them; this is particularly the case for conifers. Thus, if maintenance administrators knew the spatial position of plants along the railway system, it may be feasible to mechanically harvest them. Primary data were collected outdoors comprising around 700 leaves and conifer seedlings from 11 species. These were then photographed in a laboratory environment. In order to classify the species in the acquired image set, a machine learning approach known as Bag-of-Features (BoF) was chosen. Irrespective of the chosen type of feature extraction and classifier, the ability to classify a previously unseen plant correctly was greater than 85%. The maintenance planning of vegetation control could be improved if plants were recognised and localised. It may be feasible to mechanically harvest them (in particular, woody plants). In addition, listed endangered species growing on the trackbeds can be avoided. Both cases are likely to reduce the amount of herbicides, which often is in the interest of public opinion. Bearing in mind that natural objects like plants are often more heterogeneous within their own class rather than outside it, the results do indeed present a stable classification performance, which is a sound prerequisite in order to later take the next step to include a natural background. Where relevant, species can also be listed under the Endangered Species Act.

The teaching of reading English in a foreign language in Libyan universities: methods and models

This study focuses on the learning and teaching of Reading in English as a Foreign Language (REFL), in Libya. The study draws on an action research process in which I sought to look critically at students and teachers of English as a Foreign Language (EFL) in Libya as they learned and taught REFL in four Libyan research sites. The Libyan EFL educational system is influenced by two main factors: the method of teaching the Holy-Quran and the long-time ban on teaching EFL by the former Libyan regime under Muammar Gaddafi. Both of these factors have affected the learning and teaching of REFL and I outline these contextual factors in the first chapter of the thesis. This investigation, and the exploration of the challenges that Libyan university students encounter in their REFL, is supported by attention to reading models. These models helped to provide an analytical framework and starting point for understanding the many processes involved in reading for meaning and in reading to satisfy teacher instructions. The theoretical framework I adopted was based, mainly and initially, on top-down, bottom-up, interactive and compensatory interactive models. I drew on these models with a view to understanding whether and how the processes of reading described in the models could be applied to the reading of EFL students and whether these models could help me to better understand what was going on in REFL. The diagnosis stage of the study provided initial data collected from four Libyan research sites with research tools including video-recorded classroom observations, semi-structured interviews with teachers before and after lesson observation, and think-aloud protocols (TAPs) with 24 students (six from each university) in which I examined their REFL reading behaviours and strategies. This stage indicated that the majority of students shared behaviours such as reading aloud, reading each word in the text, articulating the phonemes and syllables of words, or skipping words if they could not pronounce them. Overall this first stage indicated that alternative methods of teaching REFL were needed in order to encourage ‘reading for meaning’ that might be based on strategies related to eventual interactive reading models adapted for REFL. The second phase of this research project was an Intervention Phase involving two team-teaching sessions in one of the four stage one universities. In each session, I worked with the teacher of one group to introduce an alternative method of REFL. This method was based on teaching different reading strategies to encourage the students to work towards an eventual interactive way of reading for meaning. A focus group discussion and TAPs followed the lessons with six students in order to discuss the 'new' method. Next were two video-recorded classroom observations which were followed by an audio-recorded discussion with the teacher about these methods. Finally, I conducted a Skype interview with the class teacher at the end of the semester to discuss any changes he had made in his teaching or had observed in his students' reading with respect to reading behaviour strategies, and reactions and performance of the students as he continued to use the 'new' method. The results of the intervention stage indicate that the teacher, perhaps not surprisingly, can play an important role in adding to students’ knowledge and confidence and in improving their REFL strategies. For example, after the intervention stage, students began to think about the title, and to use their own background knowledge to comprehend the text. The students employed, also, linguistic strategies such as decoding and, above all, the students abandoned the behaviour of reading for pronunciation in favour of reading for meaning. Despite the apparent efficacy of the alternative method, there are, inevitably, limitations related to the small-scale nature of the study and the time I had available to conduct the research. There are challenges, too, related to the students’ first language, the idiosyncrasies of the English language, the teacher training and continuing professional development of teachers, and the continuing political instability of Libya. The students’ lack of vocabulary and their difficulties with grammatical functions such as phrasal and prepositional verbs, forms which do not exist in Arabic, mean that REFL will always be challenging. Given such constraints, the ‘new’ methods I trialled and propose for adoption can only go so far in addressing students’ difficulties in REFL. Overall, the study indicates that the Libyan educational system is underdeveloped and under resourced with respect to REFL. My data indicates that the teacher participants have received little to no professional developmental that could help them improve their teaching in REFL and skills in teaching EFL. These circumstances, along with the perennial problem of large but varying class sizes; student, teacher and assessment expectations; and limited and often poor quality resources, affect the way EFL students learn to read in English. Against this background, the thesis concludes by offering tentative conclusions; reflections on the study, including a discussion of its limitations, and possible recommendations designed to improve REFL learning and teaching in Libyan universities.

Machine learning per l'idrologia: applicazione del metodo random forests per la previsione degli eventi di piena fluviale con un approccio probabilistico

Le tecniche di Machine Learning sono molto utili in quanto consento di massimizzare l’utilizzo delle informazioni in tempo reale. Il metodo Random Forests può essere annoverato tra le tecniche di Machine Learning più recenti e performanti. Sfruttando le caratteristiche e le potenzialità di questo metodo, la presente tesi di dottorato affronta due casi di studio differenti; grazie ai quali è stato possibile elaborare due differenti modelli previsionali. Il primo caso di studio si è incentrato sui principali fiumi della regione Emilia-Romagna, caratterizzati da tempi di risposta molto brevi. La scelta di questi fiumi non è stata casuale: negli ultimi anni, infatti, in detti bacini si sono verificati diversi eventi di piena, in gran parte di tipo “flash flood”. Il secondo caso di studio riguarda le sezioni principali del fiume Po, dove il tempo di propagazione dell’onda di piena è maggiore rispetto ai corsi d’acqua del primo caso di studio analizzato. Partendo da una grande quantità di dati, il primo passo è stato selezionare e definire i dati in ingresso in funzione degli obiettivi da raggiungere, per entrambi i casi studio. Per l’elaborazione del modello relativo ai fiumi dell’Emilia-Romagna, sono stati presi in considerazione esclusivamente i dati osservati; a differenza del bacino del fiume Po in cui ai dati osservati sono stati affiancati anche i dati di previsione provenienti dalla catena modellistica Mike11 NAM/HD. Sfruttando una delle principali caratteristiche del metodo Random Forests, è stata stimata una probabilità di accadimento: questo aspetto è fondamentale sia nella fase tecnica che in fase decisionale per qualsiasi attività di intervento di protezione civile. L'elaborazione dei dati e i dati sviluppati sono stati effettuati in ambiente R. Al termine della fase di validazione, gli incoraggianti risultati ottenuti hanno permesso di inserire il modello sviluppato nel primo caso studio all’interno dell’architettura operativa di FEWS.

Left orthotopic lung transplant in rats: the learning process.

INTRODUCTION: The orthotopic left lung transplantation model in rats has been developed to answer a variety of scientific questions in transplant immunology and in the related fields of respiratory diseases. However, its widespread use has been hampered by the complexity of the procedure. AIM OF THE RESEARCH: Our purpose is to provide a detailed description of the procedure of this technique, including the complications and difficulties from the very first microsurgical step until the ultimate successful completion of the transplant procedure. MATERIALS AND METHODS: The transplant procedures were performed by two collaborating transplant surgeons with microsurgical and thoracic surgery skills. A total of 150 left lung transplants in rats were performed. Twenty-seven syngeneic (Lewis to Lewis) and 123 allogeneic (Brown-Norway to Lewis) lung transplants were performed using the cuff technique. RESULTS: In first 50 transplant procedures, post-transplant survival rate was 74% of which 54% reached the end-point of 3 or 7 days post-transplant; whole complication rate was 66%. In the subsequent 50 transplant surgeries (from 51 to 100) post-transplant survival rate increased to 88% of which 56% reached the end-point; whole complication rate was 32 %. In the final 50 transplants (from 101 to 150) post-transplant survival rate was confirmed to be 88% of which 74% reached the end-point; whole complication rate was again 32 %. CONCLUSIONS: One hundred-fifty transplants can represent a reasonable number of procedures to obtain a satisfactory surgical outcome. Training period with simpler animal models is mandatory to develop anesthesiological and microsurgical skills required for successfully develop this model. The collaboration between at least two microsurgeons is mandatory to perform all the simultaneous procedures required for completing the transplant surgery.

Reinforcement learning over encrypted data

Reinforcement learning is a particular paradigm of machine learning that, recently, has proved times and times again to be a very effective and powerful approach. On the other hand, cryptography usually takes the opposite direction. While machine learning aims at analyzing data, cryptography aims at maintaining its privacy by hiding such data. However, the two techniques can be jointly used to create privacy preserving models, able to make inferences on the data without leaking sensitive information. Despite the numerous amount of studies performed on machine learning and cryptography, reinforcement learning in particular has never been applied to such cases before. Being able to successfully make use of reinforcement learning in an encrypted scenario would allow us to create an agent that efficiently controls a system without providing it with full knowledge of the environment it is operating in, leading the way to many possible use cases. Therefore, we have decided to apply the reinforcement learning paradigm to encrypted data. In this project we have applied one of the most well-known reinforcement learning algorithms, called Deep Q-Learning, to simple simulated environments and studied how the encryption affects the training performance of the agent, in order to see if it is still able to learn how to behave even when the input data is no longer readable by humans. The results of this work highlight that the agent is still able to learn with no issues whatsoever in small state spaces with non-secure encryptions, like AES in ECB mode. For fixed environments, it is also able to reach a suboptimal solution even in the presence of secure modes, like AES in CBC mode, showing a significant improvement with respect to a random agent; however, its ability to generalize in stochastic environments or big state spaces suffers greatly.

Machine Learning methods for hepatocellular malignancies segmentation and MVI prediction

The aim of this thesis project is to automatically localize HCC tumors in the human liver and subsequently predict if the tumor will undergo microvascular infiltration (MVI), the initial stage of metastasis development. The input data for the work have been partially supplied by Sant'Orsola Hospital and partially downloaded from online medical databases. Two Unet models have been implemented for the automatic segmentation of the livers and the HCC malignancies within it. The segmentation models have been evaluated with the Intersection-over-Union and the Dice Coefficient metrics. The outcomes obtained for the liver automatic segmentation are quite good (IOU = 0.82; DC = 0.35); the outcomes obtained for the tumor automatic segmentation (IOU = 0.35; DC = 0.46) are, instead, affected by some limitations: it can be state that the algorithm is almost always able to detect the location of the tumor, but it tends to underestimate its dimensions. The purpose is to achieve the CT images of the HCC tumors, necessary for features extraction. The 14 Haralick features calculated from the 3D-GLCM, the 120 Radiomic features and the patients' clinical information are collected to build a dataset of 153 features. Now, the goal is to build a model able to discriminate, based on the features given, the tumors that will undergo MVI and those that will not. This task can be seen as a classification problem: each tumor needs to be classified either as “MVI positive” or “MVI negative”. Techniques for features selection are implemented to identify the most descriptive features for the problem at hand and then, a set of classification models are trained and compared. Among all, the models with the best performances (around 80-84% ± 8-15%) result to be the XGBoost Classifier, the SDG Classifier and the Logist Regression models (without penalization and with Lasso, Ridge or Elastic Net penalization).

Deep Learning Architectures for time of arrival detection in Acoustic Emissions Monitoring

Acoustic Emission (AE) monitoring can be used to detect the presence of damage as well as determine its location in Structural Health Monitoring (SHM) applications. Information on the time difference of the signal generated by the damage event arriving at different sensors is essential in performing localization. This makes the time of arrival (ToA) an important piece of information to retrieve from the AE signal. Generally, this is determined using statistical methods such as the Akaike Information Criterion (AIC) which is particularly prone to errors in the presence of noise. And given that the structures of interest are surrounded with harsh environments, a way to accurately estimate the arrival time in such noisy scenarios is of particular interest. In this work, two new methods are presented to estimate the arrival times of AE signals which are based on Machine Learning. Inspired by great results in the field, two models are presented which are Deep Learning models - a subset of machine learning. They are based on Convolutional Neural Network (CNN) and Capsule Neural Network (CapsNet). The primary advantage of such models is that they do not require the user to pre-define selected features but only require raw data to be given and the models establish non-linear relationships between the inputs and outputs. The performance of the models is evaluated using AE signals generated by a custom ray-tracing algorithm by propagating them on an aluminium plate and compared to AIC. It was found that the relative error in estimation on the test set was < 5% for the models compared to around 45% of AIC. The testing process was further continued by preparing an experimental setup and acquiring real AE signals to test on. Similar performances were observed where the two models not only outperform AIC by more than a magnitude in their average errors but also they were shown to be a lot more robust as compared to AIC which fails in the presence of noise.

Machine Learning Unsupervised Methods in the Design of an On-board Health Monitoring System for Satellite Applications

The dissertation starts by providing a description of the phenomena related to the increasing importance recently acquired by satellite applications. The spread of such technology comes with implications, such as an increase in maintenance cost, from which derives the interest in developing advanced techniques that favor an augmented autonomy of spacecrafts in health monitoring. Machine learning techniques are widely employed to lay a foundation for effective systems specialized in fault detection by examining telemetry data. Telemetry consists of a considerable amount of information; therefore, the adopted algorithms must be able to handle multivariate data while facing the limitations imposed by on-board hardware features. In the framework of outlier detection, the dissertation addresses the topic of unsupervised machine learning methods. In the unsupervised scenario, lack of prior knowledge of the data behavior is assumed. In the specific, two models are brought to attention, namely Local Outlier Factor and One-Class Support Vector Machines. Their performances are compared in terms of both the achieved prediction accuracy and the equivalent computational cost. Both models are trained and tested upon the same sets of time series data in a variety of settings, finalized at gaining insights on the effect of the increase in dimensionality. The obtained results allow to claim that both models, combined with a proper tuning of their characteristic parameters, successfully comply with the role of outlier detectors in multivariate time series data. Nevertheless, under this specific context, Local Outlier Factor results to be outperforming One-Class SVM, in that it proves to be more stable over a wider range of input parameter values. This property is especially valuable in unsupervised learning since it suggests that the model is keen to adapting to unforeseen patterns.

Novel Machine-Learning Based IC Testing Strategy

The final goal of the thesis should be a real-world application in the production test data environment. This includes the pre-processing of the data, building models and visualizing the results. To do this, different machine learning models, outlier prediction oriented, should be investigated using a real dataset. Finally, the different outlier prediction algorithms should be compared, and their performance discussed.

Collaborative technologies, higher order thinking and self-sufficient learning: A case study of adult learners

The inclusion of online elements in learning environments is becoming commonplace in Post Compulsory Education. A variety of research into the value of such elements is available, and this study aims to add further evidence by looking specifically at the use of collaborative technologies such as online discussion forums and wikis to encourage higher order thinking and self-sufficient learning. In particular, the research examines existing pedagogical models including Salmon’s five-stage model, along with other relevant literature. A case study of adult learners in community-based learning centres forms the basis of the research, and as a result of the findings, an arrow model is suggested as a framework for online collaboration that emphasises the learner, mentions pre-course preparation and then includes three main phases of activity: post, interact and critique. This builds on Salmon’s five-stage model and has the benefit of being flexible and responsive, as well as allowing for further development beyond the model, particularly in a blended learning environment.

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.

Development of advanced methods for the simulation of the reacting mixture formation in internal combustion engines with the use of machine learning algorithms

Besides increasing the share of electric and hybrid vehicles, in order to comply with more stringent environmental protection limitations, in the mid-term the auto industry must improve the efficiency of the internal combustion engine and the well to wheel efficiency of the employed fuel. To achieve this target, a deeper knowledge of the phenomena that influence the mixture formation and the chemical reactions involving new synthetic fuel components is mandatory, but complex and time intensive to perform purely by experimentation. Therefore, numerical simulations play an important role in this development process, but their use can be effective only if they can be considered accurate enough to capture these variations. The most relevant models necessary for the simulation of the reacting mixture formation and successive chemical reactions have been investigated in the present work, with a critical approach, in order to provide instruments to define the most suitable approaches also in the industrial context, which is limited by time constraints and budget evaluations. To overcome these limitations, new methodologies have been developed to conjugate detailed and simplified modelling techniques for the phenomena involving chemical reactions and mixture formation in non-traditional conditions (e.g. water injection, biofuels etc.). Thanks to the large use of machine learning and deep learning algorithms, several applications have been revised or implemented, with the target of reducing the computing time of some traditional tasks by orders of magnitude. Finally, a complete workflow leveraging these new models has been defined and used for evaluating the effects of different surrogate formulations of the same experimental fuel on a proof-of-concept GDI engine model.

A take on complexity: bio-molecules and human metabolism interaction modelling for health and nutrition with machine learning

The advent of omic data production has opened many new perspectives in the quest for modelling complexity in biophysical systems. With the capability of characterizing a complex organism through the patterns of its molecular states, observed at different levels through various omics, a new paradigm of investigation is arising. In this thesis, we investigate the links between perturbations of the human organism, described as the ensemble of crosstalk of its molecular states, and health. Machine learning plays a key role within this picture, both in omic data analysis and model building. We propose and discuss different frameworks developed by the author using machine learning for data reduction, integration, projection on latent features, pattern analysis, classification and clustering of omic data, with a focus on 1H NMR metabolomic spectral data. The aim is to link different levels of omic observations of molecular states, from nanoscale to macroscale, to study perturbations such as diseases and diet interpreted as changes in molecular patterns. The first part of this work focuses on the fingerprinting of diseases, linking cellular and systemic metabolomics with genomic to asses and predict the downstream of perturbations all the way down to the enzymatic network. The second part is a set of frameworks and models, developed with 1H NMR metabolomic at its core, to study the exposure of the human organism to diet and food intake in its full complexity, from epidemiological data analysis to molecular characterization of food structure.

Methods for integrating machine learning and constrained optimization

In the framework of industrial problems, the application of Constrained Optimization is known to have overall very good modeling capability and performance and stands as one of the most powerful, explored, and exploited tool to address prescriptive tasks. The number of applications is huge, ranging from logistics to transportation, packing, production, telecommunication, scheduling, and much more. The main reason behind this success is to be found in the remarkable effort put in the last decades by the OR community to develop realistic models and devise exact or approximate methods to solve the largest variety of constrained or combinatorial optimization problems, together with the spread of computational power and easily accessible OR software and resources. On the other hand, the technological advancements lead to a data wealth never seen before and increasingly push towards methods able to extract useful knowledge from them; among the data-driven methods, Machine Learning techniques appear to be one of the most promising, thanks to its successes in domains like Image Recognition, Natural Language Processes and playing games, but also the amount of research involved. The purpose of the present research is to study how Machine Learning and Constrained Optimization can be used together to achieve systems able to leverage the strengths of both methods: this would open the way to exploiting decades of research on resolution techniques for COPs and constructing models able to adapt and learn from available data. In the first part of this work, we survey the existing techniques and classify them according to the type, method, or scope of the integration; subsequently, we introduce a novel and general algorithm devised to inject knowledge into learning models through constraints, Moving Target. In the last part of the thesis, two applications stemming from real-world projects and done in collaboration with Optit will be presented.