Learning Methods and Algorithms for Semantic Text Classification across Multiple Domains

Information is nowadays a key resource: machine learning and data mining techniques have been developed to extract high-level information from great amounts of data. As most data comes in form of unstructured text in natural languages, research on text mining is currently very active and dealing with practical problems. Among these, text categorization deals with the automatic organization of large quantities of documents in priorly defined taxonomies of topic categories, possibly arranged in large hierarchies. In commonly proposed machine learning approaches, classifiers are automatically trained from pre-labeled documents: they can perform very accurate classification, but often require a consistent training set and notable computational effort. Methods for cross-domain text categorization have been proposed, allowing to leverage a set of labeled documents of one domain to classify those of another one. Most methods use advanced statistical techniques, usually involving tuning of parameters. A first contribution presented here is a method based on nearest centroid classification, where profiles of categories are generated from the known domain and then iteratively adapted to the unknown one. Despite being conceptually simple and having easily tuned parameters, this method achieves state-of-the-art accuracy in most benchmark datasets with fast running times. A second, deeper contribution involves the design of a domain-independent model to distinguish the degree and type of relatedness between arbitrary documents and topics, inferred from the different types of semantic relationships between respective representative words, identified by specific search algorithms. The application of this model is tested on both flat and hierarchical text categorization, where it potentially allows the efficient addition of new categories during classification. Results show that classification accuracy still requires improvements, but models generated from one domain are shown to be effectively able to be reused in a different one.

Cascading events triggering industrial accidents: quantitative assessment of NaTech and domino scenarios

The so called cascading events, which lead to high-impact low-frequency scenarios are rising concern worldwide. A chain of events result in a major industrial accident with dreadful (and often unpredicted) consequences. Cascading events can be the result of the realization of an external threat, like a terrorist attack a natural disaster or of “domino effect”. During domino events the escalation of a primary accident is driven by the propagation of the primary event to nearby units, causing an overall increment of the accident severity and an increment of the risk associated to an industrial installation. Also natural disasters, like intense flooding, hurricanes, earthquake and lightning are found capable to enhance the risk of an industrial area, triggering loss of containment of hazardous materials and in major accidents. The scientific community usually refers to those accidents as “NaTechs”: natural events triggering industrial accidents. In this document, a state of the art of available approaches to the modelling, assessment, prevention and management of domino and NaTech events is described. On the other hand, the relevant work carried out during past studies still needs to be consolidated and completed, in order to be applicable in a real industrial framework. New methodologies, developed during my research activity, aimed at the quantitative assessment of domino and NaTech accidents are presented. The tools and methods provided within this very study had the aim to assist the progress toward a consolidated and universal methodology for the assessment and prevention of cascading events, contributing to enhance safety and sustainability of the chemical and process industry.

A Web GIS Decision Support System For Coastal Risk Assessment and Mitigation Planning

Coastal flooding poses serious threats to coastal areas around the world, billions of dollars in damage to property and infrastructure, and threatens the lives of millions of people. Therefore, disaster management and risk assessment aims at detecting vulnerability and capacities in order to reduce coastal flood disaster risk. In particular, non-specialized researchers, emergency management personnel, and land use planners require an accurate, inexpensive method to determine and map risk associated with storm surge events and long-term sea level rise associated with climate change. This study contributes to the spatially evaluation and mapping of social-economic-environmental vulnerability and risk at sub-national scale through the development of appropriate tools and methods successfully embedded in a Web-GIS Decision Support System. A new set of raster-based models were studied and developed in order to be easily implemented in the Web-GIS framework with the purpose to quickly assess and map flood hazards characteristics, damage and vulnerability in a Multi-criteria approach. The Web-GIS DSS is developed recurring to open source software and programming language and its main peculiarity is to be available and usable by coastal managers and land use planners without requiring high scientific background in hydraulic engineering. The effectiveness of the system in the coastal risk assessment is evaluated trough its application to a real case study.

Internazionalizzazione e traduzione: Un'esperienza presso l'azienda TR Turoni S.R.L.

This dissertation was conducted within the project Language Toolkit, which has the aim of integrating the worlds of work and university. In particular, it consists of the translation into English of documents commissioned by the Italian company TR Turoni and its primary purpose is to demonstrate that, in the field of translation for companies, the existing translation support tools and software can optimise and facilitate the translation process. The work consists of five chapters. The first introduces the Language Toolkit project, the TR Turoni company and its relationship with the CERMAC export consortium. After outlining the current state of company internationalisation, the importance of professional translators in enhancing the competitiveness of companies that enter new international markets is highlighted. Chapter two provides an overview of the texts to be translated, focusing on the textual function and typology and on the addressees. After that, manual translation and the main software developed specifically for translators are described, with a focus on computer-assisted translation (CAT) and machine translation (MT). The third chapter presents the target texts and the corresponding translations. Chapter four is dedicated to the analysis of the translation process. The first two texts were translated manually, with the support of a purpose-built specialized corpus. The following two documents were translated with the software SDL Trados Studio 2011 and its applications. The last texts were submitted to the Google Translate service and to a process of pre and post-editing. Finally, in chapter five conclusions are drawn about the main limits and potentialities of the different translations techniques. In addition to this, the importance of an integrated use of all available instruments is underlined.

Proposta di traduzione commentata di un articolo scientifico sulla sperimentazione animale: risorse, metodi e terminologia.

The aim of this dissertation is to provide a translation from English into Italian of a highly specialized scientific article published by the online journal ALTEX. In this text, the authors propose a roadmap for how to overcome the acknowledged scientific gaps for the full replacement of systemic toxicity testing using animals. The main reasons behind this particular choice are my personal interest in specialized translation of scientific texts and in the alternatives to animal testing. Moreover, this translation has been directly requested by the Italian molecular biologist and clinical biochemist Candida Nastrucci. It was not possible to translate the whole article in this project, for this reason, I decided to translate only the introduction, the chapter about skin sensitization, and the conclusion. I intend to use the resources that were created for this project to translate the rest of the article in the near future. In this study, I will show how a translator can translate such a specialized text with the help of a field expert using CAT Tools and a specialized corpus. I will also discuss whether machine translation can prove useful to translate this type of document. This work is divided into six chapters. The first one introduces the main topic of the article and explains my reasons for choosing this text; the second one contains an analysis of the text type, focusing on the differences and similarities between Italian and English conventions. The third chapter provides a description of the resources that were used to translate this text, i.e. the corpus and the CAT Tools. The fourth one contains the actual translation, side-by-side with the original text, while the fifth one provides a general comment on the translation difficulties, an analysis of my translation choices and strategies, and a comment about the relationship between the field expert and the translator. Finally, the last chapter shows whether machine translation and post-editing can be an advantageous strategy to translate this type of document. The project also contains two appendixes. The first one includes 54 complex terminological sheets, while the second one includes 188 simple terminological sheets.

Deep learning for computer vision: a comparison between convolutional neural networks and hierarchical temporal memories on object recognition tasks

In recent years, Deep Learning techniques have shown to perform well on a large variety of problems both in Computer Vision and Natural Language Processing, reaching and often surpassing the state of the art on many tasks. The rise of deep learning is also revolutionizing the entire field of Machine Learning and Pattern Recognition pushing forward the concepts of automatic feature extraction and unsupervised learning in general. However, despite the strong success both in science and business, deep learning has its own limitations. It is often questioned if such techniques are only some kind of brute-force statistical approaches and if they can only work in the context of High Performance Computing with tons of data. Another important question is whether they are really biologically inspired, as claimed in certain cases, and if they can scale well in terms of "intelligence". The dissertation is focused on trying to answer these key questions in the context of Computer Vision and, in particular, Object Recognition, a task that has been heavily revolutionized by recent advances in the field. Practically speaking, these answers are based on an exhaustive comparison between two, very different, deep learning techniques on the aforementioned task: Convolutional Neural Network (CNN) and Hierarchical Temporal memory (HTM). They stand for two different approaches and points of view within the big hat of deep learning and are the best choices to understand and point out strengths and weaknesses of each of them. CNN is considered one of the most classic and powerful supervised methods used today in machine learning and pattern recognition, especially in object recognition. CNNs are well received and accepted by the scientific community and are already deployed in large corporation like Google and Facebook for solving face recognition and image auto-tagging problems. HTM, on the other hand, is known as a new emerging paradigm and a new meanly-unsupervised method, that is more biologically inspired. It tries to gain more insights from the computational neuroscience community in order to incorporate concepts like time, context and attention during the learning process which are typical of the human brain. In the end, the thesis is supposed to prove that in certain cases, with a lower quantity of data, HTM can outperform CNN.

Transport of Functionally Appropriate Tools by Capuchin Monkeys (Cebus apella)

Capuchin monkeys are notable among New World monkeys for their widespread use of tools. They use both hammer tools and insertion tools in the wild to acquire food that would be unobtainable otherwise. Evidence indicates that capuchins transport stones to anvil sites and use the most functionally efficient stones to crack nuts. We investigated capuchins’ assessment of functionality by testing their ability to select a tool that was appropriate for two different tool-use tasks: A stone for a hammer task and a stick for an insertion task. To select the appropriate tools, the monkeys investigated a baited tool-use apparatus (insertion or hammer), traveled to a location in their enclosure where they could no longer see the apparatus, made a selection between two tools (stick or stone), and then could transport the tool back to the apparatus to obtain a walnut. Four capuchins were first trained to select and use the appropriate tool for each apparatus. After training, they were then tested by allowing them to view a baited apparatus and then travel to a location 8 m distant where they could select a tool while out of view of the apparatus. All four monkeys chose the correct tool significantly more than expected and transported the tools back to the apparatus. Results confirm capuchins’ propensity for transporting tools, demonstrate their capacity to select the functionally appropriate tool for two different tool-use tasks, and indicate that they can retain the memory of the correct choice during a travel time of several seconds.

Extensive clonal spread and extreme longevity in saw palmetto, a foundation clonal plant

The lack of effective tools have hampered our ability to assess the size, growth and ages of clonal plants. With Serenoa repens (saw palmetto) as a model, we introduce a novel analytical framework that integrates DNA fingerprinting and mathematical modelling to simulate growth and estimate ages of clonal plants. We also demonstrate the application of such life-history information of clonal plants to provide insight into management plans. Serenoa is an ecologically important foundation species in many Southeastern United States ecosystems; yet, many land managers consider Serenoa a troublesome invasive plant. Accordingly, management plans have been developed to reduce or eliminate Serenoa with little understanding of its life history. Using Amplified Fragment Length Polymorphisms, we genotyped 263 Serenoa and 134 Sabal etonia (a sympatric non-clonal palmetto) samples collected from a 20 X 20 m study plot in Florida scrub. Sabal samples were used to assign small field-unidentifiable palmettos to Serenoa or Sabal and also as a negative control for clone detection. We then mathematically modelled clonal networks to estimate genet ages. Our results suggest that Serenoa predominantly propagate via vegetative sprouts and 10000-year-old genets may be common, while showing no evidence of clone formation by Sabal. The results of this and our previous studies suggest that: (i) Serenoa has been part of scrub associations for thousands of years, (ii) Serenoa invasion are unlikely and (ii) once Serenoa is eliminated from local communities, its restoration will be difficult. Reevaluation of the current management tools and plans is an urgent task.

Extensive clonal spread and extreme longevity in saw palmetto, a foundation clonal plant

The lack of effective tools has hampered our ability to assess the size, growth and ages of clonal plants. With Serenoa repens (saw palmetto) as a model, we introduce a novel analytical frame work that integrates DNA fingerprinting and mathematical modelling to simulate growth and estimate ages of clonal plants. We also demonstrate the application of such life-history information of clonal plants to provide insight into management plans. Serenoa is an ecologically important foundation species in many Southeastern United States ecosystems; yet, many land managers consider Serenoa a troublesome invasive plant. Accordingly, management plans have been developed to reduce or eliminate Serenoa with little understanding of its life history. Using Amplified Fragment Length Polymorphisms, we genotyped 263 Serenoa and 134 Sabal etonia (a sympatric non-clonal palmetto) samples collected from a 20 x 20 m study plot in Florida scrub. Sabal samples were used to assign small field-unidentifiable palmettos to Serenoa or Sabal and also as a negative control for clone detection. We then mathematically modelled clonal networks to estimate genet ages. Our results suggest that Serenoa predominantly propagate via vegetative sprouts and 10000-year-old genets maybe common, while showing no evidence of clone formation by Sabal. The results of this and our previous studies suggest that: (i) Serenoa has been part of scrub associations for thousands of years, (ii) Serenoa invasions are unlikely and (ii) once Serenoa is eliminated from local communities, its restoration will be difficult. Reevaluation of the current management tools and plans is an urgent task.

Research for sustainable development: Foundations, experiences, and perspectives

Research for Sustainable Development is based on the experiences of a decade of inter- and transdisciplinary research in partnership in nine regions of the world. It presents 29 articles in which interdisciplinary teams reflect on the foundations of sustainability-oriented research, propose and illustrate concrete concepts, tools, and approaches to overcome the challenges of such research, and show how research practice related to specific issues of sustainable development has led to new thematic and methodological insights. The book seeks to stimulate the advancement of research towards more relevant, scientifically sound, and concrete contributions to realising the vision of sustainable development.

The Design of XML-Based Model and Experiment Description Languages for Network Simulation

The Simulation Automation Framework for Experiments (SAFE) is a project created to raise the level of abstraction in network simulation tools and thereby address issues that undermine credibility. SAFE incorporates best practices in network simulationto automate the experimental process and to guide users in the development of sound scientific studies using the popular ns-3 network simulator. My contributions to the SAFE project: the design of two XML-based languages called NEDL (ns-3 Experiment Description Language) and NSTL (ns-3 Script Templating Language), which facilitate the description of experiments and network simulationmodels, respectively. The languages provide a foundation for the construction of better interfaces between the user and the ns-3 simulator. They also provide input to a mechanism which automates the execution of network simulation experiments. Additionally,this thesis demonstrates that one can develop tools to generate ns-3 scripts in Python or C++ automatically from NSTL model descriptions.

Generalized eMC implementation for Monte Carlo dose calculation of electron beams from different machine types

The electron Monte Carlo (eMC) dose calculation algorithm available in the Eclipse treatment planning system (Varian Medical Systems) is based on the macro MC method and uses a beam model applicable to Varian linear accelerators. This leads to limitations in accuracy if eMC is applied to non-Varian machines. In this work eMC is generalized to also allow accurate dose calculations for electron beams from Elekta and Siemens accelerators. First, changes made in the previous study to use eMC for low electron beam energies of Varian accelerators are applied. Then, a generalized beam model is developed using a main electron source and a main photon source representing electrons and photons from the scattering foil, respectively, an edge source of electrons, a transmission source of photons and a line source of electrons and photons representing the particles from the scrapers or inserts and head scatter radiation. Regarding the macro MC dose calculation algorithm, the transport code of the secondary particles is improved. The macro MC dose calculations are validated with corresponding dose calculations using EGSnrc in homogeneous and inhomogeneous phantoms. The validation of the generalized eMC is carried out by comparing calculated and measured dose distributions in water for Varian, Elekta and Siemens machines for a variety of beam energies, applicator sizes and SSDs. The comparisons are performed in units of cGy per MU. Overall, a general agreement between calculated and measured dose distributions for all machine types and all combinations of parameters investigated is found to be within 2% or 2 mm. The results of the dose comparisons suggest that the generalized eMC is now suitable to calculate dose distributions for Varian, Elekta and Siemens linear accelerators with sufficient accuracy in the range of the investigated combinations of beam energies, applicator sizes and SSDs.

Securities: Its Conception and Protection of Interests of its Holders

Theoretical studies of the problems of the securities markets in the Russian Federation incline to one or other of the two traditional approaches. The first consists of comparing the definition of "valuable paper" set forth in the current legislation of the Russian Federation, with the theoretical model of "Wertpapiere" elaborated by German scholars more than 90 years ago. The problem with this approach is, in Mr. Pentsov's opinion, that any new features of the definition of "security" that do not coincide with the theoretical model of "Wertpapiere" (such as valuable papers existing in non-material, electronic form) are claimed to be incorrect and removed from the current legislation of the Russian Federation. The second approach works on the basis of the differentiation between the Common Law concept of "security" and the Civil Law concept of "valuable paper". Mr. Pentsov's research, presented in an article written in English, uses both methodological tools and involves, firstly, a historical study of the origin and development of certain legal phenomena (securities) as they evolved in different countries, and secondly, a comparative, synchronic study of equivalent legal phenomena as they exist in different countries today. Employing the first method, Mr. Pentsov divided the historical development of the conception of "valuable paper" in Russia into five major stages. He found that, despite the existence of a relatively wide circulation of valuable papers, especially in the second half of the 19th century, Russian legislation before 1917 (the first stage) did not have a unified definition of valuable paper. The term was used, in both theoretical studies and legislation, but it covered a broad range of financial instruments such as stocks, bonds, government bonds, promissory notes, bills of exchange, etc. During the second stage, also, the legislation of the USSR did not have a unified definition of "valuable paper". After the end of the "new economic policy" (1922 - 1930) the stock exchanges and the securities markets in the USSR, with a very few exceptions, were abolished. And thus during the third stage (up to 1985), the use of valuable papers in practice was reduced to foreign economic relations (bills of exchange, stocks in enterprises outside the USSR) and to state bonds. Not surprisingly, there was still no unified definition of "valuable paper". After the beginning of Gorbachev's perestroika, a securities market began to re-appear in the USSR. However, the successful development of securities markets in the USSR was retarded by the absence of an appropriate regulatory framework. The first effort to improve the situation was the adoption of the Regulations on Valuable Papers, approved by resolution No. 590 of the Council of Ministers of the USSR, dated June 19, 1990. Section 1 of the Regulation contained the first statutory definition of "valuable paper" in the history of Russia. At the very beginning of the period of transition to a market economy, a number of acts contained different definitions of "valuable paper". This diversity clearly undermined the stability of the Russian securities market and did not achieve the goal of protecting the investor. The lack of unified criteria for the consideration of such non-standard financial instruments as "valuable papers" significantly contributed to the appearance of numerous fraudulent "pyramid" schemes that were outside of the regulatory scheme of Russia legislation. The situation was substantially improved by the adoption of the new Civil Code of the Russian Federation. According to Section 1 of Article 142 of the Civil Code, a valuable paper is a document that confirms, in compliance with an established form and mandatory requisites, certain material rights whose realisation or transfer are possible only in the process of its presentation. Finally, the recent Federal law No. 39 - FZ "On the Valuable Papers Market", dated April 22 1996, has also introduced the term "emission valuable papers". According to Article 2 of this Law, an "emission valuable paper" is any valuable paper, including non-documentary, that simultaneously has the following features: it fixes the composition of material and non-material rights that are subject to confirmation, cession and unconditional realisation in compliance with the form and procedure established by this federal law; it is placed by issues; and it has equal amount and time of realisation of rights within the same issue regardless of when the valuable paper was purchased. Thus the introduction of the conception of "emission valuable paper" became the starting point in the Russian federation's legislation for the differentiation between the legal regimes of "commercial papers" and "investment papers" similar to the Common Law approach. Moving now to the synchronic, comparative method of research, Mr. Pentsov notes that there are currently three major conceptions of "security" and, correspondingly, three approaches to its legal definition: the Common Law concept, the continental law concept, and the concept employed by Japanese Law. Mr. Pentsov proceeds to analyse the differences and similarities of all three, concluding that though the concept of "security" in the Common Law system substantially differs from that of "valuable paper" in the Continental Law system, nevertheless the two concepts are developing in similar directions. He predicts that in the foreseeable future the existing differences between these two concepts will become less and less significant. On the basis of his research, Mr. Pentsov arrived at the conclusion that the concept of "security" (and its equivalents) is not a static one. On the contrary, it is in the process of permanent evolution that reflects the introduction of new financial instruments onto the capital markets. He believes that the scope of the statutory definition of "security" plays an extremely important role in the protection of investors. While passing the Securities Act of 1933, the United States Congress determined that the best way to achieve the goal of protecting investors was to define the term "security" in sufficiently broad and general terms so as to include within the definition the many types of instruments that in the commercial world fall within the ordinary concept of "security' and to cover the countless and various devices used by those who seek to use the money of others on the promise of profits. On the other hand, the very limited scope of the current definition of "emission valuable paper" in the Federal Law of the Russian Federation entitled "On the Valuable Papers Market" does not allow the anti-fraud provisions of this law to be implemented in an efficient way. Consequently, there is no basis for the protection of investors. Mr. Pentsov proposes amendments which he believes would enable the Russian markets to become more efficient and attractive for both foreign and domestic investors.

Analyzing, Capturing and Taming Software Change

Software systems need to continuously change to remain useful. Change appears in several forms and needs to be accommodated at different levels. We propose ChangeBoxes as a mechanism to encapsulate, manage, analyze and exploit changes to software systems. Our thesis is that only by making change explicit and manipulable can we enable the software developer to manage software change more effectively than is currently possible. Furthermore we argue that we need new insights into assessing the impact of changes and we need to provide new tools and techniques to manage them. We report on the results of some initial prototyping efforts, and we outline a series of research activities that we have started to explore the potential of ChangeBoxes.