1731 research outputs found
Sort by
Network and Cascade Representation Learning: Algorithms based on Information Diffusion Events
Full text linkNetwork representation learning (NRL) and cascade representation learn- ing (CRL) are fundamental backbones of different kinds of network analysis problems. They are usually carried out in settings where the structure of the network under consideration is known. Motivated by real-world prob- lems, this study presents several algorithms for scenarios where the network structure is partially or completely unknown. The objective of network representation learning is to identify a mapping function that projects sparse and high-dimensional network graphs into a dense latent representation, which preserves the original information about nodes and their neighborhoods. The notion of neighborhood, however, be- comes illusive when the network structure is partially or completely hidden.
Inspired by previous results, in our thesis work we have developed novel algorithms that are resilient to such lack of knowledge. These results estab- lish a correlation between the properties of the network and different kind of node activities performed over it, information which is generally more available and can be easily observed. In particular, we focus on diffusion events – also called cascades – such as shares, retweets and hashtags.
In the first of our contributions, we have developed a novel NRL algorithm called Mineral, a simple technique that combines the observed cascades with the partially accessible network structure by sampling artificial cas- cades. Node representation is then learned from the observed and sampled cascades by using the SkipGram model that is widely used for word representation learning in natural language documents.
In our second contribution, called NetTensor, we assume that the network structure is completely hidden and we propose novel techniques that are capable to estimate both the hidden neighborhood (proximity) and the similarity of nodes. Such estimated values are then used to learn a unified embedding of nodes using a scalable truncated singular value decomposition and deep autoencoders.
In addition to the NRL algorithms, we have also proposed a novel CRL algorithm called cas2vec for virality (popularity) prediction. Again, we pursue a network-agnostic approach following the above assumption that the network structure is completely unknown. Unlike prior studies that rely on manual feature extraction, cas2vec automatically learns cascade representations based on convolutional neural networks, that are effective in predicting virality of cascades.
We have carried out extensive experiments using several real-world datasets for all of our methods and compared them against strong baselines from the state-of-the-art, achieving significantly better results than many of them
Monarquía Hispánica, governo del territorio e attività pastorale nella diocesi di Lima (1580-1606)
Full text linkL’intento di questa ricerca è di definire l’importanza del sistema di circolazione delle informazioni per il governo della Monarquía Hispánica, focalizzando l’attenzione sulla diocesi di Lima negli anni di governo dell’arcivescovo Toribio di Mogrovejo (1580-1606). All’interno di questo contesto più ampio, si vuole anche mostrare l’influenza delle osservazioni e dei suggerimenti inviati da Mogrovejo al re all’interno della legislazione indiana, in particolare all’interno delle reales cedulas, e della successiva Recopilación de Leyes de Indias
Constitutional Balancing and Fundamental Labour Rights: an Analytical Approach to the Italian and Spanish Case Law on Post-Crisis Reforms
Full text linkA number of labour and social security reforms adopted by EU countries in the aftermath of the 2008 crisis have been challenged before national Constitutional Courts, which have sought to resolve the conflict between fundamental labour rights and conflicting economic interests, through balancing exercises. The thesis investigates the legal reasoning developed by the Italian and Spanish Constitutional Courts in crucial post-crisis judgments. The study draws on neo-constitutional theories and starts from the assumption that fundamental labour rights, understood as rights aimed at protecting workers and their dignity either during their working life or after retirement, must be fully enforced, albeit can be subject to limitations in order to protect other rights and interests constitutionally guaranteed.
In order to achieve this objective, the thesis focuses on the jurisprudential theories on balancing characteristic of the two judicial constitutional traditions addressed and designs an analytical framework that allows a comprehensive assessment of the units of analysis. Overall, the research has shed light over a number of issues, which have a general relevance as far as concerns the relationship between Constitutional Courts and fundamental labour rights. With regard to the specific cases investigated, the study suggests that both Courts tend to supervise the balancing conducted by the legislator, rather than to balance actively the conflicting constitutional interests. However, the techniques applied substantially diverge. The analysis shows that while the Spanish Judge has failed to both apply the proportionality test and guarantee a full enforcement of fundamental labour rights, the unstructured and dialectic technique traditionally used by the Italian Court has been – more – functional to this aim. On the other hand, in both cases, despite the significant differences, the Courts have uphold the limitations to the scope of fundamental labour rights, imposed by post-crisis policies
Identification, signaling and exploitation of social preferences: An experimental analysis
Full text linkThe Doctoral dissertation centers on social preferences. Three experimental studies address the identification methods and the implications of distributional preference types and pro-sociality on economic decision-making.
In Chapter 1, the identification of social preference types using distributive choices is discussed. A thorough review shows that the two main approaches - parametric and non-parametric- have been productively used but produced inconsistent results in previous studies. The experiment in this Chapter is designed to examine the categorical agreement between the two methods: whether they classify the same subject into the same type.
Chapter 2 presents a laboratory experiment investigating whether people strategically signal a certain type of social preferences. I consider four different distributional types and compare the distribution of these types under two settings: with and without strategic reasoning.
In Chapter 3, I report an experimental study on the strategic exploitation of others' pro-sociality for own's benefit. This study is conducted within the principal-agent framework and aims to test whether employers make use of workers' pro-social motivation, offering a compensation scheme which is tailored to workers' pro-sociality
Learning Morphology for Open-Vocabulary Neural Machine Translation
No full textState-of-the-art neural machine translation systems typically have low accuracy in translating rare or unseen words due to the requirement of using a fixed-size word vocabulary during training. In addition to controlling the model complexity, this limitation is also related to the difficulty of learning accurate word representations under conditions of high data sparsity. This problem is an important bottleneck on performance, especially in morphologically-rich languages, where the word vocabulary tends to be huge and sparse. In this dissertation, we propose to solve the vocabulary limitation problem in neural machine translation by integrating morphology learning within the translation model, aiding to learn richer word representations in terms of phonological and morphological information. Our model improves the accuracy while translating into low-resource and morphologically-rich languages and shows better generalization capability over varieties of languages with different morphological characteristics
La rappresentazione della Grande Guerra nella letteratura inglese contemporanea.
Full text linkLa tesi esamina cinque romanzi storici sulla Grande Guerra pubblicati in Inghilterra tra il 2010 e il 2015: My Dear I Wanted to Tell You e The Heroes' Welcome di Louisa Young, Toby's Room di Pat Barker, The Lie di Helen Dunmore e Field Service di Robert Edric. L'analisi dei romanzi permette di riflettere su come il conflitto viene immaginato e rappresentato alle soglie del Centenario e, in un'ottica più ampia, offre l'occasione per ragionare sulla natura della narrazione storica e sul rapporto tra storia e finzione
An eye tracking exploration of cognitive reflection in consumer decision-making
Full text linkThe works presented in this thesis are the result of the experiments conducted in the Cognitive and Experimental Economics Laboratory (CEEL) and in the Consumer Neuroscience Laboratory (NCLab) of the Economics and Management Department at the University of Trento. The aim of this research is to study the influence of cognitive impulsivity on commercial problem-solving and consumer decision-making. We focused on the attentional aspects related to the decision-making process as analyzed by the eye movements. The first section will present the main topic of the thesis, the key tool used to conduct the experiments (eye tracker) and the three papers; the latter will compose the second, third and fourth chapter. All the chapters have a common thread: to shed light on the cognitive aspects of problem-solving and their implications for the consumer decision-making process as analyzed through gaze behaviour.
- The aim of the first paper, “The role of numeracy, cognitive reflection and attentional patterns in commercial problem-solving” by Dorigoni, Polonio, Graffeo and Bonini, is to analyze the predictive power of two important cognitive abilities, numeracy and cognitive reflection, in two different problem solving scenarios with high numerical components.
- The aim of the second paper, “Getting the best deal: Effects of cognitive reflection on mental accounting of choice attributes” by Dorigoni, Cadonna and Bonini is to understand if people with low cognitive reflection are more prone to mental accounting across attributes of the same product; low cognitive reflectors do not integrate all the attribute costs and consequently they do not always choose the best deal.
- The aim of the third paper, “Cognitive reflection and gaze behaviour in visual tasks” by Dorigoni, Rajsic and Bonini is to demonstrate that cognitive reflection has predictive power on heuristics and biases related to perceptual and visual tasks. This result is extremely important because it reflects a different disposition to see and analyze the information depending on the cognitive impulsivity
Choice-supportive misremembering: A robust phenomenon?
Full text linkAlthough the literature on choice-supportive memory – remembering the items of chosen options as more preferential and those of non-chosen options as less preferential than they actually were – is scarce and scattered, it has widely been accepted as a solid phenomenon that could lead to biased future decisions. In the published studies to date, different types of such misremembering have been observed using rather dissimilar methods, with the large majority testing memory solely with source recognition. The characteristics of the material to be remembered have not been particularly varied and the effect of different delay levels has not been properly investigated. Thus, at the onset of this project, there was a lack of insight into the nature and robustness of the phenomenon and no systematic review of the relevant literature had been done to provide an integrative assessment of its status.
The objectives and scope of the current project are the following: (1) to conduct a systematic literature review on the phenomenon of choice-supportive misremembering; (2) to propose a new comprehensive taxonomy of the different types of memory distortions after choice, and (3) to investigate the empirical support for the proposed taxonomy and explore the conditions necessary for the choice-supportive misremembering effect.
The first experiment involved four choice scenarios and had a typical design for studies in the decision-making literature (information presented in tables with a low number of attributes and only the value of each one differing between the two options). Memory was tested with free and cued recall only, and the delay between the choice and the memory tests was manipulated on three levels (2 minutes, 20 minutes, and 2 days). The results fully supported the proposed taxonomy, but also highlighted the absence of choice-supportive misremembering despite the high statistical power of the tests.
In the second experiment, the scenarios and options were more complex and verbose, with a higher number of items. Four different scenarios were used and presented in two different formats (‘narrative’ vs list) as well as in an alignable and an unalignable version. The narrative versions presented the items of the two options in blocks of text (Option A above Option B), whereas the list versions displayed each item on a separate line (Option A adjacent to Option B). In the alignable version, the items presented had corresponding items on the same dimension in both options, while there was no such correspondence in the unalignable version. This time, memory was tested with free and cued recall in addition to a source recognition test that has hitherto been more typical for studies on choice-supportive misremembering. In common with the first study, this second experiment provided full support for the proposed taxonomy, but it also demonstrated the existence of the choice-supportive misremembering effect regardless of the manipulated variables.
Considered in unity, the results of the two experiments suggest that the type of stimuli used is a decisive factor and confirm that the phenomenon does not occur with the kind of materials typically used in the decision-making literature. Indeed, when participants are more likely to compare the options item by item rather than rely on gist-based processing, they do not seem to exhibit choice-supportive memory effects. This observation and the findings from the literature review and the two experiments are discussed in light of the various theories that have been proposed to explain misremembering. Finally, suggestions on how to further expand the knowledge through new research studies on specific populations are put forward
ERP correlates of semantic and syntactic processing in cochlear implant users.
Full text linkDeafness is a common sensory loss condition that affects a relatively large portion of the population around the world. Profound sensorineural deafness is the condition that we will consider in this work given its compatibility with cochlear implantation. Cochlear implant (CI) is a neuroprosthesis that allows the acoustic signal to bypass the impaired inner hear system, by sending an electrical input directly to the acoustic nerve. Although it represents an important success among neuroprosthesis, it still represents a matter of interest in the field of cognitive neuroscience. Specifically, it is still unclear how it impacts language processing in prelingually and postlingually deaf CI users where the presence and the nature of a critical period for language acquisition has been long debated (Chapter 1). To this aim, we selected well-established ERP components that have been linked to language processing like the N400 and the P600: the first usually is elicited by semantic incongruities and the second is elicited in response to syntactic manipulations (Chapter 2). We adopted a widely employed EEG paradigm where participants had to read a large sample of sentences one word at a time and we also tested subjects with a set of behavioral tasks. Sentences were semantically incongruent or contained a subject-verb number-agreement violation; an equal number of correct sentences were used as a control condition. We tested CI recipients with early deafness onset (3 years) as well as a large population of normal hearing participants that ranged from 12 up to 70 years old. The first study of this thesis will compare ERP components of language processing between preverbal CI users and a group of age-matched hearing controls. Results revealed comparable P600 and a larger N400 in CI users as well as interesting correlations between components and age at implantation that support the existent general consensus in favor of early implantation (Chapter 5). The second study employed the same RSVP paradigm and compared postlingually deaf CI participants with a sample of age peers. This study revealed that, despite the implanted group seems to have similar overall results to NH controls, CI can still impact language processing as revealed by the correlation between the N400 effect and the use of the CI (Chapter 6). Lastly, we analyzed the sample of control participants to better understand how aging can impact ERP components of language processing. We observed that both the N400 and the P600 (but also other components) are modulated by aging (Chapter 4).
Together these results proved that the study of language and its neural signatures (ERPs) is of primary interest, not only for clinical purposes but also for the advancement of psycholinguistic knowledge
Prosecutorial Discretion and its Judicial Review at the International Criminal Court: A Practice-based Analysis of the Relationship between the Prosecutor and Judges
Full text linkThe permanent system of international criminal justice created through the Rome Statute envisages a wide margin of discretion for prosecutorial action, under the constraint of various forms of judicial supervision. Nevertheless, legal texts provide only very limited guidance to the Office of the Prosecution and judges as to the concrete exercise of these powers and responsibilities. For this reason, prosecutorial and judicial dynamic practice plays a fundamental creative role in integrating—and sometimes transforming—the ICC static legal framework.
The present research has aimed at analysing the patterns of prosecutorial and judicial practice at the pre-trial stage of the proceedings of the ICC, with a view to comparing the law in the books and the law in action in this area of crucial importance for the legitimacy of the Court. The hypothesis that in this field there are areas of interpretive agreement (smooth relationship) and disagreement (open clash) between the relevant actors, as well as a certain degree of dissociation between the textual formant and the prosecutorial/judicial formant has been tested against the relevant practice. These empirical phenomena have then been assessed as to their possible institutional causes and (potentially detrimental) consequences, with a view to proposing institutional, procedural, administrative and legislative adjustments that may help fostering the predictability and consistency of the system.
The conclusion is that practice in this field is a fundamental test-bench for the institutional functioning of the ICC, and that it is still in the process of establishing— by means of the interplay between the OTP and judges—a satisfactory balance among the conflicting needs of flexibility and predictability; one that only pragmatic interpretive compromises can bring about in the future