"The capacity of contrastive analysis to successfully predict the errors of second language learners - a case study"

The aim of this dissertation is to show the power of contrastive analysis in successfully predicting the errors a language learner will make by means of a concrete case study. First, there is a description of what language transfer is and why it is important in the matter of second language acquisition. Second, a brief explanation of the history and development of contrastive analysis will be offered. Third, the focus of the thesis will move to an analysis of errors usually made by language learners. To conclude, the dissertation will focus on the concrete case study of a Russian learner of English: after an analysis of the errors the student is likely to make, a recorded conversation will be examined.

Self-supervised information retrieval: a novel approach based on Deep Metric Learning and Neural Language Models

Most of the existing open-source search engines, utilize keyword or tf-idf based techniques to find relevant documents and web pages relative to an input query. Although these methods, with the help of a page rank or knowledge graphs, proved to be effective in some cases, they often fail to retrieve relevant instances for more complicated queries that would require a semantic understanding to be exploited. In this Thesis, a self-supervised information retrieval system based on transformers is employed to build a semantic search engine over the library of Gruppo Maggioli company. Semantic search or search with meaning can refer to an understanding of the query, instead of simply finding words matches and, in general, it represents knowledge in a way suitable for retrieval. We chose to investigate a new self-supervised strategy to handle the training of unlabeled data based on the creation of pairs of ’artificial’ queries and the respective positive passages. We claim that by removing the reliance on labeled data, we may use the large volume of unlabeled material on the web without being limited to languages or domains where labeled data is abundant.

Structure learning of graphical models for task-oriented robot grasping

In the collective imaginaries a robot is a human like machine as any androids in science fiction. However the type of robots that you will encounter most frequently are machinery that do work that is too dangerous, boring or onerous. Most of the robots in the world are of this type. They can be found in auto, medical, manufacturing and space industries. Therefore a robot is a system that contains sensors, control systems, manipulators, power supplies and software all working together to perform a task. The development and use of such a system is an active area of research and one of the main problems is the development of interaction skills with the surrounding environment, which include the ability to grasp objects. To perform this task the robot needs to sense the environment and acquire the object informations, physical attributes that may influence a grasp. Humans can solve this grasping problem easily due to their past experiences, that is why many researchers are approaching it from a machine learning perspective finding grasp of an object using information of already known objects. But humans can select the best grasp amongst a vast repertoire not only considering the physical attributes of the object to grasp but even to obtain a certain effect. This is why in our case the study in the area of robot manipulation is focused on grasping and integrating symbolic tasks with data gained through sensors. The learning model is based on Bayesian Network to encode the statistical dependencies between the data collected by the sensors and the symbolic task. This data representation has several advantages. It allows to take into account the uncertainty of the real world, allowing to deal with sensor noise, encodes notion of causality and provides an unified network for learning. Since the network is actually implemented and based on the human expert knowledge, it is very interesting to implement an automated method to learn the structure as in the future more tasks and object features can be introduced and a complex network design based only on human expert knowledge can become unreliable. Since structure learning algorithms presents some weaknesses, the goal of this thesis is to analyze real data used in the network modeled by the human expert, implement a feasible structure learning approach and compare the results with the network designed by the expert in order to possibly enhance it.

"Portugal Vale A Pena" Un ritratto senza tempo del portogallo attraverso il prisma dei giornalisti

This dissertation aims at enhancing the cultural and linguistic skills in Portuguese of the author of this work, as it is a third language. This activity is carried out starting from the analysis and research of topics mentioned in a number of texts within a particular literary work, “Portugal Vale A Pena”. In this work, many Portuguese personalities express their vision on their country and state why Portugal matters. Since these texts have many cultural references, it can be assumed that such work of analysis and research can lead to better linguistic skills as well as a greater knowledge of the Portuguese culture. All of the chosen texts were originally written by journalists. This choice originates from the important service these professionals provide to the public as well as from the special role their work has played in my interpreting studies over the last couple of years. Chapter 1 explains why I chose this dissertation and who are the target users of these texts. Chapter 2 focuses on the role of journalists. A brief history of modern journalism is presented and its functions are analysed. This chapter also includes a section that examines which values make an event newsworthy. Attention is then paid to the evolution of Portuguese journalism, from Salazar's dictatorship until today. Teaching of journalism in Portuguese academia is also presented. Then, a selection of Portuguese-language media is offered. Chapter 3 focuses on some aspects of Portugal, with particular attention to its history. Finally, Chapter 4 presents a selection of texts from the original book. This part provides a biography of the authors, a translation with a comment and a presentation of some of the topics from the texts. To conclude, a glossary with words and expressions from the original text is included and their translations into Italian, Spanish and English are provided.

End-to-end Deep Metric Learning con Vision-Language Model per il Fashion Image Captioning

L'image captioning è un task di machine learning che consiste nella generazione di una didascalia, o caption, che descriva le caratteristiche di un'immagine data in input. Questo può essere applicato, ad esempio, per descrivere in dettaglio i prodotti in vendita su un sito di e-commerce, migliorando l'accessibilità del sito web e permettendo un acquisto più consapevole ai clienti con difficoltà visive. La generazione di descrizioni accurate per gli articoli di moda online è importante non solo per migliorare le esperienze di acquisto dei clienti, ma anche per aumentare le vendite online. Oltre alla necessità di presentare correttamente gli attributi degli articoli, infatti, descrivere i propri prodotti con il giusto linguaggio può contribuire a catturare l'attenzione dei clienti. In questa tesi, ci poniamo l'obiettivo di sviluppare un sistema in grado di generare una caption che descriva in modo dettagliato l'immagine di un prodotto dell'industria della moda dato in input, sia esso un capo di vestiario o un qualche tipo di accessorio. A questo proposito, negli ultimi anni molti studi hanno proposto soluzioni basate su reti convoluzionali e LSTM. In questo progetto proponiamo invece un'architettura encoder-decoder, che utilizza il modello Vision Transformer per la codifica delle immagini e GPT-2 per la generazione dei testi. Studiamo inoltre come tecniche di deep metric learning applicate in end-to-end durante l'addestramento influenzino le metriche e la qualità delle caption generate dal nostro modello.

Analysis of human posterior EEG alpha frequency as a function of implicit learning of probabilistic lateralized stimuli in a visual detection task

Alpha oscillatory activity has long been associated with perceptual and cognitive processes related to attention control. The aim of this study is to explore the task-dependent role of alpha frequency in a lateralized visuo-spatial detection task. Specifically, the thesis focuses on consolidating the scientific literature's knowledge about the role of alpha frequency in perceptual accuracy, and deepening the understanding of what determines trial-by-trial fluctuations of alpha parameters and how these fluctuations influence overall task performance. The hypotheses, confirmed empirically, were that different implicit strategies are put in place based on the task context, in order to maximize performance with optimal resource distribution (namely alpha frequency, associated positively with performance): “Lateralization” of the attentive resources towards one hemifield should be associated with higher alpha frequency difference between contralateral and ipsilateral hemisphere; “Distribution” of the attentive resources across hemifields should be associated with lower alpha frequency difference between hemispheres; These strategies, used by the participants according to their brain capabilities, have proven themselves adaptive or maladaptive depending on the different tasks to which they have been set: "Distribution" of the attentive resources seemed to be the best strategy when the distribution probability between hemifields was balanced: i.e. the neutral condition task. "Lateralization" of the attentive resources seemed to be more effective when the distribution probability between hemifields was biased towards one hemifield: i.e., the biased condition task.

Learning and accepting quantum physics. Re-analysis of a teaching proposal

Nella tesi è analizzata nel dettaglio una proposta didattica sulla Fisica Quantistica elaborata dal gruppo di ricerca in Didattica della Fisica dell’Università di Bologna, in collaborazione con il gruppo di ricerca in Fisica Teorica e con ricercatori del CNR di Bologna. La proposta è stata sperimentata in diverse classi V di Liceo scientifico e dalle sperimentazioni sono emersi casi significativi di studenti che non sono riusciti ad accettare la teoria quantistica come descrizione convincente ad affidabile della realtà fisica (casi di non accettazione), nonostante sembrassero aver capito la maggior parte degli argomenti e essersi ‘appropriati’ del percorso per come gli era stato proposto. Da questa evidenza sono state formulate due domande di ricerca: (1) qual è la natura di questa non accettazione? Rispecchia una presa di posizione epistemologica o è espressione di una mancanza di comprensione profonda? (2) Nel secondo caso, è possibile individuare precisi meccanismi cognitivi che possono ostacolare o facilitare l’accettazione della fisica quantistica? L’analisi di interviste individuali degli studenti ha permesso di mettere in luce tre principali esigenze cognitive (cognitive needs) che sembrano essere coinvolte nell’accettazione e nell’apprendimento della fisica quantistica: le esigenze di visualizzabilità, comparabilità e di ‘realtà’. I ‘cognitive needs’ sono stati quindi utilizzati come strumenti di analisi delle diverse proposte didattiche in letteratura e del percorso di Bologna, al fine di metterne in luce le criticità. Sono state infine avanzate alcune proposte per un suo miglioramento.