Domiziana Catalano

Work

AI Research Fellow

Leonardo

January 2021 - Present • Genoa, Italy

• Research Fellow in Leonardo's applied AI research lab.
• Strong focus on Computer Vision (text detection and scene text recognition).
• Natural Language Processing project about domain-specific Semantic Text Similarity and Named Entity Recognition.
• Exploitation of the computational power of davinci-1 HPC.
• Participation to NVIDIA and Intel workshops.
• Proposals writing for European calls (Horizon 2021).

IT Consultant

Indra

May 2018 - September 2018 • Rome, Italy

Consultant developing TIBCO solutions in BNL BNP Paribas headquarters in Rome.

Education

MSc in Artificial Intelligence and Robotics

Sapienza University of Rome

September 2018 - October 2020 • Rome, Italy
Master thesis on Computer Vision: Deep Learning techniques for Deepfake video detection.

BSc in Computer Science and Automatic Engineering

Sapienza University of Rome

September 2014 - May 2018 • Rome, Italy

TIBCO and CRM Training Course

Indra Academy

March 2018 - April 2018 • Rome, Italy

Projects

DeepFake detector

Design and implementation of two methods that detect deepfake videos using state-of-the-art deep learning techniques.

Scene Text Recognition

• Pytorch Lightning implementation of the AI system.
• Detection of text regions in in-the-wild images.
• Text Super Resolution of the detected ROIs to improve text recognition.

Domain-specific Semantic Text Similarity

• Pytorch implementation of the Natural Language Processing system.
• Documents classification based on semantic similarity.
• Clustering of documents based on semantic similarity.

Word Sense Disambiguation System

Keras implementation of a WSD system to assign the correct sense to a word in a sentence.

Music Recommendation System - WIP

Pytorch Lightning implementation of a music recommendation system.

Talking Chalks GitHub

• IoT system that can be used in a museum.
• It can help visitors during the tour providing information about the museum and the artworks.
• A smartband with the most suitable profile will be provided to each visitor, so that can everyone can have a customized experience.

TripPlanner

• Android app that helps you find and manage flights.
• You can choose fixed days or a window of time and it will find the cheapest price combination.
• You can buy tickets, save flights and check the weather at the destination.

Clever Weather GitHub

IoT system where a RIOT-OS app collects data from sensors and sends them to an Azure IoT Hub using MQTT-SN and MQTT protocol.

Customer service in cinema with Pepper

• Pepper robot can help customers providing info and helping them reserving seats.
• It can also collect film reviews and entertain people.

User Activity Recognition

Azure webapp that uses the accelerometer to identify the user activity and sends data to Azure IoT Hub.

About

I'm Domiziana, a computer science engineer. My main work interest right now is on artificial intelligence, in particular deep learning for computer vision, but I also work on some domain-specific Natural Language Processing projects. Moreover, lately I began to approach music recommendation world, in order to combine my passion for music with my work in AI field.

In my free time, I like listening to music, reading books and going to exhibitions.

IT Topics

Artificial Intelligence • Machine Learning • Deep Learning • Computer Vision • Generative Models • Natural Language Processing • Music Recommendation • Internet of Things • Mobile App

Programming

Python (openCV, PIL, PyTorch, Keras, TensorFlow, numpy, pandas, scikitlearn, etc.) • Java • C • Javascript • HTML • Scala • Matlab • TIBCO

Languages

Italian (native) • English (proficient)

Contact

You can reach me here

• Linkedin
• Email
• GitHub

Elements

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';