This document outlines the work completed during a roughly three-hundred-hour internship by Carlo Rosso at the University of Padua. The internship had several key objectives. First, the goal was to study the problem of sentiment analysis and explore the relevant scientific literature. Second, the task involved implementing, experimenting with, and comparing at least three state-of-the-art models using the Sentiment Penn Treebank corpus. Third, the implemented models had to be documented and uploaded to a Git repository to ensure the reproducibility of the results. During the internship, special attention was given to investigating the role of syntactic analysis and its compositional properties in understanding semantic phenomena, with a particular focus on sentiment analysis.
Il presente documento descrive il lavoro svolto durante il periodo di tirocinio, della durata di circa trecento ore, dal laureando Carlo Rosso presso l'Università degli Studi di Padova. Gli obiettivi da raggiungere sono stati molteplici. In primo luogo era richiesto di studiare il problema della sentiment analysis e della relativa letteratura scientifica. In secondo luogo era richiesta l'implementazione, sperimentare e confrontare almeno tre modelli allo stato dell'arte sul corpus Sentiment Penn Treebank. In terzo luogo era richiesto di documentare i modelli implementati e di caricarli in una repository git per garantire la riproducibilità dei risultati. Più nel particolare, durante il periodo di stage è stato indagato il ruolo dell'analisi sintattica e delle sue proprietà di composizionalità nella comprensione dei fenomeni semantici, con un focus specifico sulla sentiment analysis.
Confronto di Kernel Methods e Large Language Models nella Sentiment Analysis.
ROSSO, CARLO
2023/2024
Abstract
This document outlines the work completed during a roughly three-hundred-hour internship by Carlo Rosso at the University of Padua. The internship had several key objectives. First, the goal was to study the problem of sentiment analysis and explore the relevant scientific literature. Second, the task involved implementing, experimenting with, and comparing at least three state-of-the-art models using the Sentiment Penn Treebank corpus. Third, the implemented models had to be documented and uploaded to a Git repository to ensure the reproducibility of the results. During the internship, special attention was given to investigating the role of syntactic analysis and its compositional properties in understanding semantic phenomena, with a particular focus on sentiment analysis.File | Dimensione | Formato | |
---|---|---|---|
thesis.pdf
accesso aperto
Dimensione
2.06 MB
Formato
Adobe PDF
|
2.06 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/70978