1,721,978 research outputs found
Un problema aperto: «lo bello stilo» virgiliano di Dante
No full textIn Inf. 1, 79-80 Dante declares that Virgil is his sole source for poetry and presents himself as Virgil's follower. What does really Dante mean? There has been much discussion of exactly what the «bello stilo» is and where it is to be found in Dante's work, but it is not satisfactory. Perhaps Dante's formulation is only a surprising claim for a poetic identity, a strategy for showing himself as a new ‘classic’ poet
L'organizzazione centrale negli statuti dei partiti politici
No full textIl contributo esamina le previsioni degli Statuti dei partiti politici che nel corso del 2011 hanno costituito un gruppo parlamentare alla Camera e al Senato in tema di organizzazione centrale, metodi di deliberazione e definizione delle linee politiche. Il tutto nel quadro delle riflessioni sui partiti e le loro caratteristiche che la dottrina pubblicistica abitualmente svolge a partire dall'art. 49 Cost
Un "Fiore" tra i commenti
No full textCome nasce l’idea del commento a un poemetto duecentesco, attorno al quale aleggia il nome di Dante Alighieri? Si ripercorrerà la genesi del lavoro di curatela realizzato per gli Oscar Mondadori nel 1996, attraverso l’analisi delle componenti fondamentali: la destinazione del volume, le necessità concomitanti di scientificità e di divulgazione, le scelte finali. A diciotto anni di distanza se ne tenta un bilancio anche alla luce degli sviluppi di ricerca successivi e degli orientamenti interpretativi odierni.The article charts the reception of the annotated edition of The Flower published by Luca Carlo Rossi in the pocketbook edition of Oscar Mondadori (Milan) in 1996. An account is given of the editorial decisions, the criteria of textual selection, the intended readership, the attribution of the work to Dante and the reactions that the volume elicited. Subsequent commentaries on The Flower are then contextualized and an alternative title is proposed in place of the traditional and apocryphal one
Ethics and interdisciplinarity in computational social science.
No full textDuring the last few years a growing amount of content produced by Internet users has become publicly available online. These data come from a variety of places, including popular social web services like Facebook and Twitter, consumer services like Amazon or weblogs.The research opportunities opened up by this socio-technological innovation are, as shown by the growing literature on the topic, huge. At the same time new challenges for social scientists arise. In this paper we will focus on two of the main challenges posed to the growth of the so-called computational social science: interdisciplinarity and ethics. While the searchability and persistence of this information make it ideal for sociological research, a quantitative approach is still challenging because of the size and complexity of the data.Collecting, storing and analyzing these data often require technical skills beyond the traditional curricula of social scientists. These projects require, in fact, collaboration with computer scientists. Nevertheless developing a common interdisciplinary project is often challenging because of the different backgrounds of the researchers.At the same time the availability of this content poses a challenge concerning privacy and research ethics. Due to the amount of data and the fact that the real identity of the author is often hidden behind a nickname, it is often impossible to ask the subject involved to consent to the use of their data. On the other hand, especially in the first wave of web 2.0, this information has been – intentionally or not – publicly shared by the users. While a technique of dis-embedding the identity of the user from the content analyzed is often the solution used to bypass this issue, an even more important privacy-related challenge for computational social science is emerging. Due to the wide adoption of social network sites such Facebook or Google+, where a user may decide to share his content with his/her group of friends only, the amount of public data will change and decrease in the future. We will discuss this issue by enumerating a number of possible future scenarios
Media & Generation: how user generated content reshape generational identity in the Mass Media
No full textForme emergenti della proprietà privata e nuove tecnologia, una prospettiva sociologica. - Tesi di dottorato
No full textTesi di Dottorat
Evaluation Metrics for Data Scarcity: Assessing the Generalizability and Robustness of Generative Models for Data Synthesis
No full textIl deep learning ha raggiunto un successo senza precedenti in vari domini, come la computer vision e il natural language processing. Tuttavia, la sua efficacia è limitata dalla necessità di grandi quantità di dati di addestramento. I metodi generativi rappresentano una soluzione promettente, offrendo la capacità di creare dataset sintetici in situazioni di scarsità di dati, ma la generalizzabilità e robustezza di tali metodi rimangono incerte.
Questa tesi affronta questa lacuna, gettando le basi per futuri lavori in questo ambito. Il lavoro si contestualizza attorno a due casi di studio: lo Zero-Shot Learning (ZSL), un dominio teorico, e la previsione delle traiettorie, un'applicazione pratica.
Lo ZSL coinvolge l'addestramento di modelli su classi "viste" e la loro valutazione su classi "non viste", impiegando metodi generativi per la data augmentation.
Questa tesi introduce un framework per misurare la generalizzabilità e la robustezza dei modelli ZSL, attraverso la valutazione di tali modelli su diversi split delle classi e dello spazio semantico.
Inoltre, questa tesi dimostra come tecniche di riduzione della dimensionalità possano migliorare le prestazioni dei modelli nei dataset fine-grained.
Per la previsione di traiettorie, cruciale in campi che vanno dalla guida autonoma alla pianificazione urbana, questa tesi esplora l'uso di modelli generativi per la sintesi di dati in fase di inferenza.
I risultati indicano che le GAN producono previsioni più realistiche in scenari multimodali, dove i modelli LSTM tendono a predire un comportamento medio tra quelli possibili.
Vengono introdotte nuove metriche e dataset per valutare la generalizzabilità in contesti reali, con un focus sia sulle traiettorie umane che su quelle dei veicoli.
I risultati ottenuti dimostrano la necessità e l'efficacia dei metodi proposti per garantire che i metodi generativi siano generalizzabili e robusti nei contesti esaminati.Deep learning has achieved unprecedented success in various domains, such as computer vision and natural language processing. However, its effectiveness is constrained by the necessity for extensive training data. Generative methods are a promising solution, offering the capability to create synthetic datasets in situations of data scarcity. Nevertheless, their real-world generalizability and robustness remain uncertain, and there is a lack of good evaluation metrics to effectively assess these properties.
This thesis addresses this gap, laying the groundwork for future research into this topic. It is contextualized around two case studies: Zero-Shot Learning (ZSL), a theoretical domain, and trajectory prediction, a practical application. ZSL, which trains models on "seen" classes to work on "unseen" ones, employs generative models for data augmentation. This thesis introduces an evaluation framework for assessing the generalizability and robustness of ZSL models, through the systematic evaluation of such models on different splits of the classes and semantic space. This thesis also demonstrates how dimensionality reduction can improve model performance, particularly with fine-grained datasets.
For trajectory prediction, crucial in fields from autonomous vehicles to urban planning, this thesis explores generative models for synthesizing data during inference, comparing the effectiveness of LSTM and GAN models in different scenarios. The findings indicate that GANs produce more realistic predictions in multimodal scenarios, whereas LSTM models fall short as they tend to average out the possible behaviors. Novel metrics and datasets are introduced to assess generalizability in real-world contexts, with a focus on both human and vehicle trajectories.
The findings demonstrate the necessity and effectiveness of the proposed methods for ensuring that generative methods are generalizable and robust in the examined contexts
- …
