About me

After being struck by the charm of the Artificial Intelligence thanks to my first experience in Parsit, I continued to deepen my knowledge in this field by applying it to Natural Language Processing (NLP) with the help and the experience of Matteo Grella, until I obtained my current employment at Advanced Analytics, where I'm working on the research and the development of algorithms of Artificial Intelligence and neural networks applied to the analysis of texts in natural languages.
I've been able to include me in this complicated reality of work thanks to the skills obtained during my degree course in Computer Engineering at Politecnico di Torino.

Current working position

Software developer at Advanced Analytics GmbH.


Turin, Italy - Konstanz, Germany.

Working experiences

[2015 - now]
A german company which operates in the cognitive intelligence and provides tools which use semantic processes to analyze structured and unstructured data coming from different types of sources, like Social Media, Open Source information, Analysts and others.

[2014 - 2015]
Built from the foundations of Parsit, the technological background of Damantic offers cutting-edge text analytics software capable of understanding texts written in natural languages (Italian, English, Spanish, French and German) and tackle the Big Data challenges, offering advanced technological solutions.

An italian startup founded by Matteo Grella and Marco Nicola. It is the first initiative to give industrial and commercial value to the results of their research work in the field of Natural Language Processing (NLP).


Non-Projective Dependency Parsing via Latent Heads Representation (LHR)
[Matteo Grella and Simone Cangialosi - 2018]
Abstract: In this paper we introduce a novel approach based on a bidirectional recurrent autoencoder to perform globally optimized non-projective dependency parsing via semisupervised learning. The syntactic analysis is completed at the end of the neural process that generates a Latent Heads Representation (LHR), without any algorithmic constraint and with a linear complexity. The resulting “latent syntactic structure” can be used directly in other semantic tasks. The LHR is transformed into the usual dependency tree computing a simple vectors similarity. We believe that our model has the potential to compete with much more complex state-of-the-art parsing architectures.