Studies of CMS data access patterns with machine learning techniques

This thesis presents a study of the Grid data access patterns in distributed analysis in the CMS experiment at the LHC accelerator. This study ranges from the deep analysis of the historical patterns of access to the most relevant data types in CMS, to the exploitation of a supervised Machine Learning classification system to set-up a machinery able to eventually predict future data access patterns - i.e. the so-called dataset “popularity” of the CMS datasets on the Grid - with focus on specific data types. All the CMS workflows run on the Worldwide LHC Computing Grid (WCG) computing centers (Tiers), and in particular the distributed analysis systems sustains hundreds of users and applications submitted every day. These applications (or “jobs”) access different data types hosted on disk storage systems at a large set of WLCG Tiers. The detailed study of how this data is accessed, in terms of data types, hosting Tiers, and different time periods, allows to gain precious insight on storage occupancy over time and different access patterns, and ultimately to extract suggested actions based on this information (e.g. targetted disk clean-up and/or data replication). In this sense, the application of Machine Learning techniques allows to learn from past data and to gain predictability potential for the future CMS data access patterns. Chapter 1 provides an introduction to High Energy Physics at the LHC. Chapter 2 describes the CMS Computing Model, with special focus on the data management sector, also discussing the concept of dataset popularity. Chapter 3 describes the study of CMS data access patterns with different depth levels. Chapter 4 offers a brief introduction to basic machine learning concepts and gives an introduction to its application in CMS and discuss the results obtained by using this approach in the context of this thesis.

Detection and Grading of Prostate Cancer in Temporal Enhanced Ultrasound using Convolutional Neural Networks

Prostate cancer is the most common non-dermatological cancer amongst men in the developed world. The current definitive diagnosis is core needle biopsy guided by transrectal ultrasound. However, this method suffers from low sensitivity and specificity in detecting cancer. Recently, a new ultrasound based tissue typing approach has been proposed, known as temporal enhanced ultrasound (TeUS). In this approach, a set of temporal ultrasound frames is collected from a stationary tissue location without any intentional mechanical excitation. The main aim of this thesis is to implement a deep learning-based solution for prostate cancer detection and grading using TeUS data. In the proposed solution, convolutional neural networks are trained to extract high-level features from time domain TeUS data in temporally and spatially adjacent frames in nine in vivo prostatectomy cases. This approach avoids information loss due to feature extraction and also improves cancer detection rate. The output likelihoods of two TeUS arrangements are then combined to form our novel decision support system. This deep learning-based approach results in the area under the receiver operating characteristic curve (AUC) of 0.80 and 0.73 for prostate cancer detection and grading, respectively, in leave-one-patient-out cross-validation. Recently, multi-parametric magnetic resonance imaging (mp-MRI) has been utilized to improve detection rate of aggressive prostate cancer. In this thesis, for the first time, we present the fusion of mp-MRI and TeUS for characterization of prostate cancer to compensates the deficiencies of each image modalities and improve cancer detection rate. The results obtained using TeUS are fused with those attained using consolidated mp-MRI maps from multiple MR modalities and cancer delineations on those by multiple clinicians. The proposed fusion approach yields the AUC of 0.86 in prostate cancer detection. The outcomes of this thesis emphasize the viable potential of TeUS as a tissue typing method. Employing this ultrasound-based intervention, which is non-invasive and inexpensive, can be a valuable and practical addition to enhance the current prostate cancer detection.

‘‘Model age-based” and ‘‘copy when uncertain” biases in children’s social learning of a novel task

Theoretical models of social learning predict that individuals can benefit from using strategies that specify when and whom to copy. Here the interaction of two social learning strategies, model age-based biased copying and copy when uncertain, was investigated. Uncertainty was created via a systematic manipulation of demonstration efficacy (completeness) and efficiency (causal relevance of some actions). The participants, 4- to 6-year-old children (N = 140), viewed both an adult model and a child model, each of whom used a different tool on a novel task. They did so in a complete condition, a near-complete condition, a partial demonstration condition, or a no-demonstration condition. Half of the demonstrations in each condition incorporated causally irrelevant actions by the models. Social transmission was assessed by first responses but also through children’s continued fidelity, the hallmark of social traditions. Results revealed a bias to copy the child model both on first response and in continued interactions. Demonstration efficacy and efficiency did not affect choice of model at first response but did influence solution exploration across trials, with demonstrations containing causally irrelevant actions decreasing exploration of alternative methods. These results imply that uncertain environments can result in canalized social learning from specific classes of mode

Socratic flipped classroom : What types of questions and tasks promote learning?

Socratic questioning stresses the importance of questioning for learning. Flipped Classroom pedagogy generates a need for effective questions and tasks in order to promote active learning. This paper describes a project aimed at finding out how different kinds of questions and tasks support students’ learning in a flipped classroom context. In this study, during the flipped courses, both the questions and tasks were distributed together with video recordings. Answers and solutions were presented and discussed in seminars, with approximately 10 participating students in each seminar. Information Systems students from three flipped classroom courses at three different levels were interviewed in focus groups about their perceptions of how different kinds of questions and tasks supported their learning process. The selected courses were organized differently, with various kinds of questions and tasks. Course one included open questions that were answered and presented at the seminar. Students also solved a task and presented the solution to the group. Course two included open questions and a task. Answers and solutions were discussed at the seminars where students also reviewed each other’s answers and solutions. Course three included online single- and multiple choice questions with real-time feedback. Answers were discussed at the seminar, with the focus on any misconceptions. In this paper we categorized the questions in accordance with Wilson (2016) as factual, convergent, divergent, evaluative, or a combination of these. In all, we found that any comprehensible question that initiates a dialogue, preferably with a set of Socratic questions, is perceived as promoting learning. This is why seminars that allow such questions and discussion are effective. We found no differences between the different kinds of Socratic questions. They were seen to promote learning so long as they made students reflect and problematize the questions. To conclude, we found that questions and tasks promote learning when they are answered and solved in a process that is characterized by comprehensibility, variation, repetition and activity.

The Proces of Curriculum and Module Design

The outcome of the inductive decision -making process of the leading project management group (PMG) was the proposal to develop three modules, Human Resource Management and Knowledge Management, Quality Management and Intercultural management, each for 10 ECTS credits. As a result of the theoretical and organisational framework and analytical phase of the project, four strategies informed the development and implemen- tation of the modules: 1. Collaboration as a principle stemming from EU collaborative policy and receiving it’s expression on all implementation levels (designing the modules, modes of learning, delivering the modules, evaluation process). 2. Building on the Bologna process masters level framework to assure ap- propriate academic level of outputs. 3. Development of value -based leadership of students through transforma- tional learning in a cross -cultural setting and continual reflection of theory in practice. 4. Continual evaluation and feedback among teachers and students as a strategy to achieve a high quality programme. In the first phase of designing the modules the collaborative strategy in particular was applied, as each module was led by one university, but members from all other universities participated in the discussions and development of the mod- ules. The Bologna process masters level framework and related standards and guidelines informed the form and method of designing the modules.

The Pleasure of Learning

The evolution, from an education centered in teaching to a perspective of learning, constitutes a change of paradigm as stated by the author in this article - who invites us to use the pleasure of learning as a facilitator for the construction of the senses. The author underlines that education must advance from the non-critical repetition of content to the generation of pleasant learning, and that the search for alternatives for an education that overcomes the traditional paradigm leads to the need to promote an authentic learning pleasure.

Experimental Project on the Safety and the Optimization of the Use of Glyphosate-Based Herbicides in the Grain Sector.

INTRODUCTION: Glyphosate is the most widely applied pesticide worldwide and it is an active ingredient of all glyphosate-based herbicides (GBHs), including in the formulation “Roundup” . It is unclear if the glyphosate present in ground water can be absorbed and translocated in different parts of the pants, particularly wheat plants. This indeed represents an important aspect for productivity (being this a powerful herbicide) and organic certification of the products (the use of glyphosate is not admitted in organic farming and the ubiquitous contamination of glyphosate in water might in theory affect the level of glyphosate in the plants). Overall, epidemiological, in vivo and in vitro studies available in literature present conflicting findings on the safety of glyphosate. METHODS: The work performed for this PhD thesis aimed to experimentally test the root absorption and the eventual translocation of the glyphosate herbicide in the different parts of the wheat plant (Triticum durum) starting from ground water. Furthermore we aimed to experimentally test the effects of the exposure to GBHs at doses of glyphosate considered to be “safe”, the US ADI of 1.75 mg/kg bw/day, defined as the chronic Reference Dose (cRfD) determined by the US EPA, in in vivo models (Sprague-Dawley rats) and in vitro models (Caco2 and L929). RESULTS: All the experimental absorption studies on wheat plants performed have given negative results in terms of the presence of glyphosate or AMPA in the grain of durum wheat. On the other hand the experimental safety studies on in vitro and in vivo models highlighted different effects at doses currently considered safe for humans and with no effects in animals. CONCLUSION: Overall the integration of the findings from absorption in plants and safety studies will serve as solid evidence-base for risk assessment and productive strategies for agriculture.

Comparison of Explicit and Implicit Methods of Cross-Cultural Learning in an International Classroom

The paper addresses a gap in the literature concerning the difference between enhanced and not enhanced cross-cultural learning in an international classroom. The objective of the described research was to clarify if the environment of international classrooms could enhance cross-cultural competences significantly enough or if additional focus on cross-cultural learning as an explicit objective of learning activities would add substantially to the experience. The research question was defined as “how can a specific exercise focused on cross-cultural learning enhance the cross-cultural skills of university students in an international classroom?”. Surveys were conducted among interna- tional students in three leading Central-European Universities in Lithuania, Poland and Hungary to measure the increase of their cross-cultural competences. The Lithuanian and Polish classes were composed of international students and concentrated on International Management/Business topics (explicit method). The Hungarian survey was done in a general business class that just happened to be international in its composition (implicit method). Overall, our findings prove that the implicit method resulted in comparable, somewhat even stronger effectiveness than the explicit method. The study method included the analyses of students’ individual increases in each study dimension and construction of a compound measure to note the overall results. Our findings confirm the power of the international classroom as a stimulating environment for latent cross-cultural learning even without specific exercises focused on cross-cultural learning itself. However, the specific exercise did induce additional learning, especially related to cross-cultural awareness and communication with representatives of other cultures, even though the extent of that learning may be interpreted as underwhelming. The main conclusion from the study is that the diversity of the students engaged in a project provided an environment that supported cross-cultural learning, even without specific culture-focused reflections or exercises.

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.

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.

Analysis and Evaluation of Tiny Machine Learning applications

The aim of TinyML is to bring the capability of Machine Learning to ultra-low-power devices, typically under a milliwatt, and with this it breaks the traditional power barrier that prevents the widely distributed machine intelligence. TinyML allows greater reactivity and privacy by conducting inference on the computer and near-sensor while avoiding the energy cost associated with wireless communication, which is far higher at this scale than that of computing. In addition, TinyML’s efficiency makes a class of smart, battery-powered, always-on applications that can revolutionize the collection and processing of data in real time. This emerging field, which is the end of a lot of innovation, is ready to speed up its growth in the coming years. In this thesis, we deploy three model on a microcontroller. For the model, datasets are retrieved from an online repository and are preprocessed as per our requirement. The model is then trained on the split of preprocessed data at its best to get the most accuracy out of it. Later the trained model is converted to C language to make it possible to deploy on the microcontroller. Finally, we take step towards incorporating the model into the microcontroller by implementing and evaluating an interface for the user to utilize the microcontroller’s sensors. In our thesis, we will have 4 chapters. The first will give us an introduction of TinyML. The second chapter will help setup the TinyML Environment. The third chapter will be about a major use of TinyML in Wake Word Detection. The final chapter will deal with Gesture Recognition in TinyML.

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.

Development of an institutional framework to guide transitions into enhanced blended learning in higher education

The rapidly changing digital landscape is having a significant influence on learning and teaching. Our study assesses the response of one higher education institution (HEI) to the changing digital landscape and its transition into enhanced blended learning, which seeks to go beyond the early implementation stage to make the most effective use of online learning technologies to enhance the student experience and student learning outcomes. Evidence from a qualitative study comprising 20 semi-structured interviews, informed by a literature review, has resulted in the development of a holistic framework to guide HEIs transitioning into enhanced blended learning. The proposed framework addresses questions relating to the why (change agents), what (institutional considerations), how (organisational preparedness) and who (stakeholders) of transitions into enhanced blended learning. The involvement of all stakeholder groups is essential to a successful institutional transition into enhanced blended learning.

How do medium naturalness and personality traits shape academic achievement and perceived learning? An experimental study of face-to-face and synchronous e-learning

This controlled experiment examined how academic achievement and cognitive, emotional and social aspects of perceived learning are affected by the level of medium naturalness (face-to-face, one-way and two-way videoconferencing) and by learners’ personality traits (extroversion–introversion and emotional stability–neuroticism). The Media Naturalness Theory explains the degree of medium naturalness by comparing its characteristics to face-to-face communication, considered to be the most natural form of communication. A total of 76 participants were randomly assigned to three experimental conditions: face-to-face, one-way and two-way videoconferencing. E-learning conditions were conducted through Zoom videoconferencing, which enables natural and spontaneous communication. Findings shed light on the trade-off involved in media naturalness: one-way videoconferencing, the less natural learning condition, enhanced the cognitive aspect of perceived learning but compromised the emotional and social aspects. Regarding the impact of personality, neurotic students tended to enjoy and succeed more in face-to-face learning, whereas emotionally stable students enjoyed and succeeded in all of the learning conditions. Extroverts tended to enjoy more natural learning environments but had lower achievements in these conditions. In accordance with the ‘poor get richer’ principle, introverts enjoyed environments with a low level of medium naturalness. However, they remained focused and had higher achievements in the face-to-face learning.

Machine learning tools for protein annotation: the cases of transmembrane β-barrel and myristoylated proteins

Biology is now a “Big Data Science” thanks to technological advancements allowing the characterization of the whole macromolecular content of a cell or a collection of cells. This opens interesting perspectives, but only a small portion of this data may be experimentally characterized. From this derives the demand of accurate and efficient computational tools for automatic annotation of biological molecules. This is even more true when dealing with membrane proteins, on which my research project is focused leading to the development of two machine learning-based methods: BetAware-Deep and SVMyr. BetAware-Deep is a tool for the detection and topology prediction of transmembrane beta-barrel proteins found in Gram-negative bacteria. These proteins are involved in many biological processes and primary candidates as drug targets. BetAware-Deep exploits the combination of a deep learning framework (bidirectional long short-term memory) and a probabilistic graphical model (grammatical-restrained hidden conditional random field). Moreover, it introduced a modified formulation of the hydrophobic moment, designed to include the evolutionary information. BetAware-Deep outperformed all the available methods in topology prediction and reported high scores in the detection task. Glycine myristoylation in Eukaryotes is the binding of a myristic acid on an N-terminal glycine. SVMyr is a fast method based on support vector machines designed to predict this modification in dataset of proteomic scale. It uses as input octapeptides and exploits computational scores derived from experimental examples and mean physicochemical features. SVMyr outperformed all the available methods for co-translational myristoylation prediction. In addition, it allows (as a unique feature) the prediction of post-translational myristoylation. Both the tools here described are designed having in mind best practices for the development of machine learning-based tools outlined by the bioinformatics community. Moreover, they are made available via user-friendly web servers. All this make them valuable tools for filling the gap between sequential and annotated data.