giulio-derasmo

AstraStatsLearn-Galaxy-image-segmentation

Image segmentation of satellite images of galaxies using machine learning and deep learning tecnique, in order to identifiy galaxies.

Bayesian-Analysis-for-Game-Players-Behaviors

A MCMC Bayesian analysis versus Frequentist Analysis of Animal Crossing: New Horizons game players in-game behavior using a Multinomial Logistic Regression Model to adjust the original paper results.

Calibration-of-a-Deep-Learning-Model

This project aim to understad if a deep learning model is calibrated (average accuracy match average confident) using Reliability Diagram and perform a re-calibration by the training with Focal Loss.

Embeddings-for-Sentiment-Analysis

Using different technique to deal with Text Classification for Sentimental Analsysi from the a base TF-IDF encoding to pre-trained sentence embeddings.

Experimenting-with-modularity-in-deep-learning

The project aim to experiment implementing a modular architecture: an early-exit model and testing it using Tensorflow.

Heart-Rate-Zones-Prediction

The aim of this work was to predict the heart rate zones. To do this we applied several data transformation techniques which we then used to pull an Xgboost model.

LLM2Vec-for-ER-task

Project for the PhD Course in Entity Resolution @Sapienza

MySQL-MongoDB-HR-Database

Create an HR database, perform realistic query and evaluating the performance of those on different DBMS like MySQL and MongoDB

Reinforcement-Learning-Projects

Repository of Reinforcement Learning projects done during the course @Sapienza

Ship-Classification-Leonardo-Labs-Kaggle-Competition

Neural Networks ensemble via majority voting in order to classify ships given non-satellite images. All the models have been trained using PyTorch with pretrained weights.

giulio-derasmo/AstraStatsLearn-Galaxy-image-segmentation

Image segmentation of satellite images of galaxies using machine learning and deep learning tecnique, in order to identifiy galaxies.

giulio-derasmo/Ship-Classification-Leonardo-Labs-Kaggle-Competition

Neural Networks ensemble via majority voting in order to classify ships given non-satellite images. All the models have been trained using PyTorch with pretrained weights.

giulio-derasmo/Experimenting-with-modularity-in-deep-learning

The project aim to experiment implementing a modular architecture: an early-exit model and testing it using Tensorflow.

giulio-derasmo/LLM2Vec-for-ER-task

Project for the PhD Course in Entity Resolution @Sapienza

giulio-derasmo/Reinforcement-Learning-Projects

Repository of Reinforcement Learning projects done during the course @Sapienza

giulio-derasmo/Bayesian-Analysis-for-Game-Players-Behaviors

A MCMC Bayesian analysis versus Frequentist Analysis of Animal Crossing: New Horizons game players in-game behavior using a Multinomial Logistic Regression Model to adjust the original paper results.

giulio-derasmo/Calibration-of-a-Deep-Learning-Model

This project aim to understad if a deep learning model is calibrated (average accuracy match average confident) using Reliability Diagram and perform a re-calibration by the training with Focal Loss.

giulio-derasmo/Embeddings-for-Sentiment-Analysis

Using different technique to deal with Text Classification for Sentimental Analsysi from the a base TF-IDF encoding to pre-trained sentence embeddings.

giulio-derasmo/Exploratory-Data-Analysis

giulio-derasmo/Exploring-StackOverflow

Analysis and implemention on graph based on StackOverFlow dataset

giulio-derasmo/Heart-Rate-Zones-Prediction

The aim of this work was to predict the heart rate zones. To do this we applied several data transformation techniques which we then used to pull an Xgboost model.

giulio-derasmo/MySQL-MongoDB-HR-Database

Create an HR database, perform realistic query and evaluating the performance of those on different DBMS like MySQL and MongoDB

giulio-derasmo/Information-Retrieval-over-MyAnimeList

Web scraping of the site MyAnimeList in order to build a search engine on the anime list based on a query over the description of it.

giulio-derasmo/Page-Rank-and-Recommendation-Systems

Use the Scikit-Network for PageRank algorithms including Topic-specific PR and improve the performance of various recommendation-systems using Surprise library

giulio-derasmo/RAG-Caching

Research on caching over RAG

giulio-derasmo/Search-Engine-Evaluation-and-Near-Duplicate-Detection

Exploiting the PyTerrier library to build a Search Engine and resolve the Near Duplicate Detection tasks.