Giuseppe Masi
Student

About
I am a PhD student in Computer Science at Sapienza University of Rome, Italy. My research focuses on applying artificial intelligence techniques to financial markets. My works start with a precise mathematical formulation to lead the machine learning model in the problem to be optimised.
Currently, I am working on generating synthetic time series with underlying causal relationships (represented as causal graphs) through generative models, particularly Transformers-based Diffusion Models. Moreover, I am also developing advanced time-series analysis to detect shocks in asset prices and classify price trends.
Apart from Machine Learning-based projects, I worked on creating a Software Defined Network to manage the redirection of traffic dynamically. Moreover, I am working on a project to optimise the distributed inference of large deep learning models, taking advantage of the Lyapunov theory to solve an optimisation problem involving carbon footprint, response time, and accuracy over time.
Finally, during the Master's Degree courses, I had the opportunity to develop NLP projects, dealing with tasks such as Word Sense Disambiguation and Learning to Rank, implementing state-of-the-art models. I also worked in the field of reinforcement learning in several tasks, including the allocation of communication channels to a set of drones.
Publications
-
V. Arrigoni, G. Masi, E. Mercanti, N. Bartolini, S. Vyetrenko
“Stock Shocks Modelling and Forecasting”
In 2023 IEEE 43rd International Conference on Distributed Computing Systems Workshops (ICDCSW)
-
G. Masi, M. Prata, M. Conti, N. Bartolini, S. Vyetrenko
“On Correlated Stock Market Traces Generation”
In 2023 ACM 4th International Conference on AI in Finance (ICAIF)
-
M. Prata, G. Masi, L. Berti, V. Arrigoni, A. Coletta, I. Cannistraci, P. Velardi, S. Vyetrenko, N. Bartolini
“LOB-Based Deep Learning Models for Stock Price Trend Prediction: A Benchmark Study”
Artificial Intelligence Review 2024
-
N. Bartolini, G. Masi, M. Prata, F. Trombetti
“Patrolling Heterogeneous Targets with FANETs”
In 2024 IEEE International Conference on Computer Communications Workshops (INFOCOMW)
-
G. Masi, A. Coletta, E. Fons, S. Vyetrenko, N. Bartolini
“CATS: Causally Associated Time-Series Generation through Diffusion Models”
Under review.
-
V. Arrigoni, G. Masi, N. Bartolini
“STOP! A Solution for Sustainable and Geo-Distributed AI Inference”
Under review.
Resume
Education
Ph.D. Fellowship in Artificial Intelligence
Oct. 2022 - Current
Sapienza, University of Rome, IT
Master's Degree in Computer Science
Oct. 2020 - Oct. 2022
Sapienza, University of Rome, IT
Final grade: 110/110 cum laude - GPA: 4/4
Thesis: Adversarial Learning to Rank Transferable Text-Based Attacks to Black-Box Neural Ranking Models: WARA and WSRA
Bachelor's Degree in Computer Science
Oct. 2017 - Oct. 2020
Tor Vergata, University of Rome, IT
Final grade: 110/110 cum laude - GPA: 3.94/4
Thesis: Diffusion in the presence of ambivalent relationships: the role of negative relationships in the complexity of the problem
Awards
Research Grant by Sapienza University 2023
“Enhancing Fairness in Financial Markets - Leveraging NLP Techniques for Textual Analysis”
Research Grant by Sapienza University 2024
“SustainaML: An Architecture for Carbon-Aware Distributed Learning Policy Programmability”
Participations in Funded Projects
Research collaboration as participant in the project NGIAtlantic.eu 2022
“Vulnerability assessment and Robust Defences for Optimised Attacks in Dynamic SDNs”
Research collaboration as participant in the project activities of the JPMorgan Faculty Award
“Understanding interdependent market dynamics: vulnerabilities and opportunities”
Skills
OOP (Python, Java, C++); C
(Deep) Machine Learning: Classification-Regression tasks
NLP Pipeline
Optimization (Gurobi)
SQL, MongoDB
Blockchain Technologies: Ethereum (Solidity)
Projects
Contact
Address
Viale Regina Elena 295, Rome