I am an Italian computer scientist living in Tuscany.
I'm interested in the fields of AI and Machine Learning,
but I also love music, good designs, and learning in general.
I enjoy problem solving, and I'm always looking for something new to discover and learn from.
M.Sc. degree in Computer Science from University of Pisa.
ITAmoji 2018: Emoji Prediction via Tree Echo State Networks
Proceedings of the Sixth Evaluation Campaign of Natural Language processing and Speech Tools for Italian (EVALITA 2018), Torino, Italy
Daniele Di Sarli, Claudio Gallicchio, Alessio Micheli —
For the "ITAmoji" EVALITA 2018 competition we mainly exploit a Reservoir Computing approach to learning, with an ensemble of models for trees and sequences.
The sentences for the models of the former kind are processed by a language parser and the words are encoded by using pretrained FastText word embeddings for
the Italian language. With our method, we ranked 3rd out of 5 teams.
GFS: a Graph-based File System Enhanced with Semantic Features
International Conference on Information System and Data Mining, Charleston, SC, USA
Daniele Di Sarli, Filippo Geraci —
In this paper we describe GFS (graph-based file system), a new hybrid file system that extends the standard hierarchical organization of files with semantic features. GFS allows the user to nest semantic spaces inside the directory hierarchy leaving system folders unaltered. Semantic spaces allow customized file tagging and leverage on browsing to guide file searching.
Design and development of the prototype of a bleeding edge semantic file system at CNR
This file system, built at CNR (Italian National Research Council), allows semantic file organization while maintaining a seamless integration with the existing systems (e.g. standard Unix tools can be used for file tagging, and file managers do not require plugins or modifications to fully work with GFS).
PoC of an automated coordination system between small-scale vehicles
This proof of concept implements a detection, communication and coordination system
between small-scale cars in a crossroad without any centralized infrastructure.
Each car is able to detect other vehicles around it and communicate with them in
order to compute the precedence or to signal an emergency situation. The system
makes use of Arduino UNOs, custom hardware, and machine learning techniques.
This library allows to specify advanced rules for matching a set of intervals (in the sense of mathematical objects). The rules are then applied using linear programming techniques, to return the best match (if available) or to generate an approximation of the expected set of intervals.
Design and development of a SAAS attendance management system. The software interfaces with attendance machines to provide real-time feedback to the users, while providing a high degree of expressiveness regarding employee schedule specifications.