1731 research outputs found
Sort by
Information Retrieval from Neurophysiological Signals
Get PDFOne of the ultimate goals of neuroscience is decoding someone's intentions directly from his/her brain activities. In this thesis, we aim at pursuing this goal in different scenarios. Firstly, we show the possibility of creating a user-centric music/movie recommender system by employing neurophysiological signals. Regarding this, we employed a brain decoding paradigm in order to classify the features extracted from brain signals of participants watching movie/music video clips, into our target classes (two broad music genres and four broad movie genres). Our results provide a preliminary experimental evidence towards user-centric music/movie content retrieval by exploiting brain signals. Secondly, we addressed one of the main issue of the applications of brain decoding algorithms. Generally, the performance of such algorithms suffers from the constraint of having few and noisy samples, which is the case in most of the neuroimaging datasets. In order to overcome this limitation, we employed an adaptation paradigm in order to transfer knowledge from another domain (e.g. large-scale image domain) to the brain domain. We experimentally show that such adaptation procedure leads to improved results. We performed such adaptation pipeline on different tasks (i.e. object recognition and genre classification) using different neuroimaging modalities (i.e. fMRI, EEG, and MEG). Thirdly, we aimed at one of the fundamental goals in brain decoding which is reconstructing the external stimuli using only the brain features. Under this scenario, we show the possibility of regressing the stimuli spectrogram using time-frequency analysis of the brain signals. Finally, we conclude the thesis by summarizing our contributions and discussing the future directions and applications of our research
Development of innovative tools for multi-objective optimization of energy systems
Get PDFFrom industrial revolution to the present day, fossil fuels are the main sources for ensuring energy supply. Fossil fuel usages have negative effects on environment that are highlighted by several local or international policy initiatives at support of the big energy transition. The effects urge energy planners to integrate renewable energies into the corresponding energy systems. However, large-scale incorporation of renewable energies into the systems is difficult because of intermittent behaviors, limited availability and economic barriers. It requires intricate balancing among different energy producing resources and the syringes among all the major energy sectors. Although it is possible to evaluate a given energy scenario (complete set of parameters describing a system) by using a simulation model, however, identifying optimal energy scenarios with respect to multiple objectives is a very difficult to accomplished. In addition, no generalized optimization framework is available that can handle all major sectors of an energy system. In this regards, we propose a complete generalized framework for identifying scenarios with respect to multiple objectives. The framework is developed by coupling a multi-objective evolutionary algorithm and EnergyPLAN. The results show that the tool has the capability to handle multiple energy sectors together; moreover, a number of optimized trade-off scenarios are identified. Furthermore, several improvements are proposed to the framework for finding better-optimized scenarios in a computationally efficient way. The framework is applied on two different real-world energy system optimization problems. The results show that the framework is capable to identify optimized scenarios both by considering recent demands and by considering projected demands. The proposed framework and the corresponding improvements make it possible to provide a complete tool for policy makers for designing optimized energy scenarios. The tool can be able to handle all major energy sectors and can be applied in short and long-term energy planning
The syntactic side of Time: processing Adverb-Verb Temporal Agreement
No full textThe aim of this thesis deals with the investigation of the cognitive mechanisms underlying the online processing of adverb-verb temporal agreement, namely the coherence in temporal features between the verb and a deictic temporal adverb (e.g. YesterdayPAST I wentPAST/*will goFUT to the jazz concert), during sentence comprehension.
There are at least two reasons that make the investigation of this phenomenon interesting and challenging at the same time. The first reason is more theoretical. Differently from well-studied phenomena such as subject-verb agreement or anaphora, the nature of the adverb-verb relation is still debated in theoretical linguistics. The debated nature of the adverb-verb temporal agreement probably relies on the peculiar properties of the constituents involved in the relation. The relation between temporal adverbs and verbs has been traditionally defined anaphoric in nature, since verbs can be bound to temporal antecedents (e.g. adverb) which allow to set a reference time that the event expressed by the verb must refer to. However, other accounts have hypothesized a structural parallelism between the adverb-verb and the subject-verb relation. One question is thus whether this similarities or dissimilarities between subject-verb agreement and temporal agreement at the theoretical level can mirror a similarity/dissimilarity at the cognitive level. The second reason is more empirical and comes from past experimental literature investigating the processing of temporal agreement. Basically, experimental evidence mainly comes from ERPs studies and results are rather sparse and heterogeneous. These ERP studies report an early detection of the temporal violation (around 200 milliseconds after the stimulus onset) but qualitatively different ERP waveforms were elicited by the target word (i.e. the verb) in the different experiments (e.g. LAN, N400, right lateralized negativities). One licit question is thus where the source of heterogeneity resides and how this phenomenon can be better investigated. Given the debated theoretical and experimental past evidence on the processing of this relation, the current work was conducted (i) investigating the pattern elicited by adverb-verb temporal violations compared to other better-studied agreement phenomena, as the one between the subject and the verb, during sentence comprehension (ii) adopting only behavioral techniques since a detailed investigation of the behavioral costs elicited by a temporal violation needs to be established prior to facing the detailed neurophysiological correlates of these processes, which are known to be subject to a larger interpretive freedom with respect to reading time differences. The core of the thesis, namely six empirical studies investigating the processing of temporal violations through different designs and techniques, is preceded by a theoretical chapter which deals with the description of Tense and deictic temporal adverbs from a semantic and syntactic point of view. The main aim of the theoretical chapter is to give a brief overview of the main linguistic theories which have investigated the nature of Tense and temporal adverbs, but also a motivation for considering the syntactic interaction between Tense and temporal adverbs, which is fundamental to preserve the grammaticality of the sentence.The first set of self-paced reading studies, in Italian, addressed two main questions: how different is the processing of adverb-verb temporal agreement with respect to other better-studied phenomena such as subject-verb number agreement? Is the different configuration between the verb and the temporal adverb that has led to heterogeneous results in past experimental literature? In the second (eye-tracking) study, in Spanish, three other questions were addressed: how differently the parser deals with the processing of number, tense and (crucially) person features when encountering a violation on the inflected verb? Does the distance between the two constituents of the dependency play any role in the detection of the violation? Finally, in the third set of eye-tracking studies in English, the processing of the adverb-verb temporal relation was tested in a more complex sentential environment, namely in sentence where the temporal adverb and the verb are separated by an embedded relative clause containing a distracting temporal element. In this set of studies, several questions were addressed: how different can be the processing of adverb-verb temporal agreement at a conspicuous distance? Is the temporal adverb-verb relation sensitive to interference effects from an illicit intervener? How differently this relation behaves with respect to subject-verb agreement and anaphora during memory retrieval? All findings collected in this work provide further evidence for a differentiation in the processing of agreement mechanisms entailing a covariance of features between two constituents within a sentence. This evidence is in line with previous accounts showing a differentiation in the processing of different features (i.e. number, person) within the same relation such as subject-verb agreement (Mancini et al. 2013), and in the processing of the same feature (e.g. number) across different relations such as subject-verb agreement and anaphora (Dillon et al. 2013). This evidence can be particularly relevant for the development of a new model of sentence parsing. In fact, among mainstream models of parsing, only Construal (Frazier & Clifton, 1996) model assumes a relation-sensitive language system. However, a differentiation in the processing of agreement phenomena is not explicitly addressed in terms of feature-related properties. On the other hand, some recent accounts have proposed different processing mechanisms depending on the feature under computation (e.g. Mancini et al., 2013; Carminati, 2005) but a specific formalization of the role of different features properties within a model of parsing has not been provided yet. The second challenge that this current work tried to face was to add more complexity into the agreement configuration testing the adverb-verb agreement relation at different linear distance. The findings here collected seems to give positive evidence on the role played by word order in the processing of the adverb-verb relation, but further investigation needs to address whether other factors may play a role and whether adverb-verb agreement is the only agreement relation which is sensitive to word order. This puzzle thus opens new questions about whether the agreement relation may change even being both the relation and the feature under computation equal. The role played by linear distance in the detection of adverb-verb temporal anomalies also suggests that parsing routines are not “stagnant”, and the language system can deal with redundant information in a very dynamic fashion
Expectations of Obedience and the Development of Moral Reasoning
Get PDFThe current dissertation tackles two core aspects of moral development: the obedience to authority and the consideration of the agents’ intention in generating a moral judgment. With respect to obedience to authority, I reported that 21-month-olds are already able to distinguish between coercion by a bully (someone who prevails using physical force) and rule by a leader (someone who is spontaneously respected by subordinates), and expect subordinates to comply with the leader’s instructions, but not with the bully’s instructions. With respect to the development of intent-based judgment, I reported that between age 4 and 6 the verbal judgment of moral goodness undergoes a shift from relying on action outcomes to relying on agent’s intentions. I argue that this shift likely reflects ancillary changes occurring outside the moral domain, such as in theory of mind or executive functioning. I also reported that, later in life, a further shift occurs in moral judgment. Older adults’ judgments, compared to younger adults’ judgments, rely less on intention and more on outcomes. This intent-to-outcome shift in old age can be explained by an age-related decline in theory of mind abilities
L'anatocismo. Contributo allo studio della teoria dell'obbligazione pecuniaria.
No full textIl lavoro nasce, quanto a stimoli, dall’instabilità delle regole che governano il “bancario” per approfondire un argomento, quello degli interessi c.d. composti, che vive da sempre nell’ambiguità (rectius, dissociazione) di formanti di segno opposto. La dinamica di conflitto, di contraddizioni, di rottura fra formante legislativo e giurisprudenza disorienta la prassi e restituisce allo studioso regole operazionali, non mere declamazioni, sovente incompatibili tra loro. L’obiettivo che ci si propone è allora quello di delineare – per il medio della comparazione giuridica – l’esatto ambito applicativo del fenomeno anatocistico, sia ripercorrendo l’evoluzione storica dell’istituto nel diritto civile e commerciale, sia attraverso un confronto con la disciplina dettata in tema di obbligazioni pecuniarie e interessi. Si è così dato ampio risalto, oltre ai temi più generali dell’(in)adempimento e degli interessi monetari con le loro tradizionali partizioni (prima tra tutte, quella in compensativi e moratori), a “province” della scienza giuridica strettamente inerenti l’istituto in esame, come gli usi, la trasparenza, l’usura, ecc. Tale inquadramento, tuttavia, coglie esclusivamente alcuni profili strutturali del fenomeno in discorso. Altri non meno significativi ne risultano trascurati: la dialettica fra autonomia privata e controllo giudiziario; il ruolo della clausola madre di buona fede oggettiva; i risultati dell’applicazione delle regole di informazione, di adeguatezza, di meritevolezza e di equilibrio economico. Un diverso inquadramento, forse più coerente con il mercato del credito europeo, profila presupposti costruttivi differenti e conseguenze applicative non trascurabili
Knowledge Networks in Emerging ICT Regional Innovation Systems: An Explorative Study of the Knowledge Network of Trentino ICT Innovation System
No full textAlthough the last thirty years Regional Innovation Systems (RIS) received great attention by policy makers, only during the last decade social networks were applied in the fields of innovation and regional economics. The majority of the existing empirical studies on networks adopt a static point of view, representing a regional knowledge network at a certain point in time, while there are few recent attempts exploring the evolution of knowledge networks and the dynamics that drive it. The present work aims at covering some of the gaps in the literature, using the dataset on collaborative projects from the ICT activity in Trentino. It introduces an original multidimensional framework to analyze the knowledge flows inside, from within and towards the regional network. It also identifies the key actors inside the region and describes their role in knowledge creation and diffusion. Concerning the spatial and temporal evolution of the knowledge networks, this thesis investigates the preferences of the economic actors operating inside regional networks, in terms of shared characteristics, while it explores the dynamics developed through time by the behavior of economic agents during high and low certainty periods, contributing to the inertia and the resilience of the regional knowledge network. The present research is the first that introduces Social Network Analysis (SNA) using data on knowledge transfer from Trentino, considering the entire universe of actors involved in the regional ICT knowledge network for the last fifteen years, and allocating it to an original multidimensional framework, in order to reveal the value of the knowledge network per se, and the impact of the regional policies on the network and not on the output of the innovation process. On the spatial evolution of networks, it explores in depth the preferences of the actors of a regional knowledge network, in order to make it more solid through strong collaborations. It proves that the effect of every kind of proximity or distance is different, while it introduces the measure of relational proximity, exploring the effect of the position of an actor inside the knowledge network in relation with the rest of the actors. However, the major finding of this thesis is the introduction of the temporal aspect in the evolution of the regional knowledge network, and the exploration of the agent behavior during periods of uncertainty. The introduction in the network evolution of an external negative event, like economic crisis, allows the deduction of useful conclusions on how the actors behave in terms of trust and collaboration creation
Enhancement of wastewater and sludge treatment processes by hydrodynamic cavitation
Get PDFIn the past decades, hydrodynamic cavitation (HC) process was the subject of study by many researchers worldwide. This phenomenon was widely studied in order to understand the reason of its negative effects on hydraulic machinery such as pumps,turbines, valves, etc. Many efforts were made in order to better understand mechanisms of HC process with the main aim of preventing its generation and trying to avoid severe physical damage such as erosions, vibrations and noises. In recent years, in order to cope with a decrease in available water resources worldwide, an increasing demand of water by population in developing/developed countries and more restrictive environmental legislations on water quality, HC was increasingly used as a novel energy-efficient technique in the field of wastewaters treatment.
The main purpose of this thesis is to investigate on the effectiveness of a modified swirling-jet device called Ecowirl reactor, patented by Econovation GmbH, Germany and
produced and commercialized by Officine Parisi s.r.l., Italy. Experimental studies were carried out in order to evaluate the effects of different operative conditions and parameters such as reactor geometry, flow rate, flow velocity, pressure, medium pH, medium concentration and medium temperature on (i) the degradation of a toxic and
carcinogenic pollutant dye (Rhodamine B, RhB) in waste dye aqueous solutions and on (ii) the improvement of activated sludge solubilisation and aerobic sludge biodegradability in the field of biological wastewater treatments.
In order to better understand the fluid dynamics into Ecowirl reactor, it was modelled. The model based on previous experimental data was implemented in a Computational Fluid Dynamics software (ANSYS, 16.2)
Planar copper containing anode-supported solid oxide fuel cells
No full textPlanar copper-containing anode supported Intermediate Temperature Solid Oxide Fuel Cells (IT-SOFCs) were produced by single step cosintering. The anode and the electrolyte were realized through water-based tape casting, the cathode being added by screen printing. A 5 mol% of Lithium oxide addition allowed reducing the Gadolinia-Doped Ceria (GDC)-based electrolyte sintering temperature below the copper oxide melting point. IT-SOFCs sintered at 950°C revealed a power density peak of 26 mW cm-2 at 650°C in H2, the maximum CuO amount within the anodic cermet being limited at 35 vol%. To improve the cell performance, the anode electrolyte thickness ratio was increased, in order to take advantage by the compressive tensile state induced by the supporting to the thinner layer, this leading to a further sintering temperature reduction and to avoid cracks due to the thermal expansion coefficient (TEC) mismatch existing between anode and electrolyte. IT-SOFCs at 900°C showed a power density peak of 200 mW cm-2 at 700°C in H2. Electronic impedance spectroscopy pointed out anode performances comparable with those obtained by using conventional Ni-based cermet electrodes. In biogas, 45 vol% CuO containing SOFC achieved a power density peak of 38 mW cm-2 at 700°C
Labour Market entry in context.Institutions, social inequalities and the early occupational Careers in the Europe
No full textThe thesis examines the process of transition from education to employment in comparative perspective. More specifically, it investigates how market and educational institutions influence the school-to-work transition processes in the European context. A comparative perspective is adopted, in order to disentangle theoretically – by means of examples – and parametrically – by means of statistical modelling – the institutional influence on the first stages of occupational careers of young school leavers. In order to provide a comprehensive view of the school-to-work transition, the thesis focuses on both the process of entry into the first employment and the occupational progression beyond that point, with a particular emphasis on the emergence of social inequalities in the early career stages
Controlling the effect of crowd noisy annotations in NLP Tasks
Get PDFNatural Language Processing (NLP) is a sub-field of Artificial Intelligence and Linguistics, with the aim of studying problems in the automatic generation and understanding of natural language. It involves
identifying and exploiting linguistic rules and variation with code to translate unstructured language data into information with a schema. Empirical methods in NLP employ machine learning techniques to automatically extract linguistic knowledge from big textual data instead
of hard-coding the necessary knowledge. Such intelligent machines require input data to be prepared in such a way that the computer can more easily find patterns and inferences. This is feasible by adding relevant metadata
to a dataset. Any metadata tag used to mark up elements of the dataset is called an annotation over the input.
In order for the algorithms to learn efficiently and effectively, the annotation done on the data must be accurate, and relevant to the task the machine is being asked to perform. In other words, the supervised machine learning methods intrinsically can not handle the inaccurate and noisy annotations and the performance of the learners have a high correlation with the quality of the input data
labels. Hence, the annotations have to be prepared by experts. However, collecting labels for large dataset is impractical to perform by a small group of qualified experts or when the experts are unavailable. This is special crucial for the recent deep learning methods which the algorithms
are starving for big supervised data. Crowdsourcing has emerged as a new paradigm for obtaining labels for
training machine learning models inexpensively and for high level of data volume. The rationale behind this concept is to harness the “wisdom of the crowd” where groups of people pool their abilities to show collective intelligence.
Although crowdsourcing is cheap and fast but collecting high quality data from the non-expert crowd requires careful attention to the task quality control management. The quality control process consists of selection of
appropriately qualified workers, providing a clear instruction or training that are understandable to non-experts and performing sanitation on the results to reduce the noise in annotations or eliminate low quality workers.
This thesis is dedicated to control the effect of crowd noisy annotations use for training the machine learning models in variety of natural language processing tasks namely: relation extraction, question answering and
recognizing textual entailment. The first part of the thesis deals with design a benchmark for evaluation Distant Supervision (DS) for relation extraction task. We propose a
baseline which involves training a simple yet accurate one-vs-all strategy using SVM classifier. Moreover, we exploit automatic feature extraction technique using convolutional tree kernels and study several example filtering techniques for improving the quality of the DS output.
In the second part, we focused on the problem of the crowd noisy annotations in training two important NLP tasks, i.e., question answering and recognizing textual entailment. We propose two learning methods to handle the noisy labels by (i) taking into account the disagreement between
crowd annotators as well as their skills for weighting instances in learning algorithms; and (ii) learning an automatic label selection model based on combining annotators characteristic and the task syntactic structure
representation as features in a joint manner.
Finally, we observe that in fine-grained tasks like relation extraction where the annotators need to have some deeper expertise, training the crowd workers has more impact on the results than simply filter-out the low quality crowd workers. Training crowd workers often requires high-quality
labeled data (namely, gold standard) to provide the instruction and feedback to the crowd workers.
We conversely, introduce a self-training strategy for crowd workers where the training examples are automatically selected via a classifier. Our study shows that even without using any gold standard, we still can train workers
which open doors toward inexpensive crowd training procedure for different NLP tasks