Brevetto depositato per la mia startup YouAreMyGuide – GhostWriter.AI:
Method for dynamic management of an itinerary, and corresponding computer system and product

A method for dynamic management of an itinerary, comprising the steps providing a first database (10) containing information on the preferences of a user, thus generating a user profile, providing a second database (20) containing travel information for each itinerary, comprising the places to visit in each itinerary, starting date and time and date and time of arrival for each itinerary, the bookings and the latest positions of the user. A route is identified, on the basis of processing of the information regarding the places to visit (20) and the user’s preferences (10). Moreover provided is a third database (30) for storing raw information made available by different data sources (A) present on the web. The processed information (B) is stored in a fourth database (40), by selecting from the raw information the data that present a high reliability. The method provides: – calculating the impact of the processed information on the route identified; – generating an itinerary; and – pre-arranging bookings on the basis of the itinerary generated. Moreover provided is an itinerary- modification module (H), which, on the basis of recognition, from the processed information (40), of a new event that has an impact on the itinerary generated, calculates a plurality of alternative routes taking into account the user’s preferences and notifies these alternative routes to the user and stores in the second database (20) the possible changes of itinerary selected by the user.

Leggi qui le rivendicazioni


Le ricerche su dblp computer science bibliography
Il mio profilo su Research Gate e quello su ORCID

ghostwriter19 @ ATE_ABSITA: Zero-Shot and ONNX to Speed up BERT on Sentiment Analysis Tasks at EVALITA 2020

With the arrival of BERT in 2018, NLP research has taken a significant step forward. However, the necessary computing power has grown accordingly.
Various distillation and optimization systems have been adopted but are costly in terms of cost-benefit ratio. The most important improvements are obtained by creating increasingly complex models with more layers and parameters.
In this research, we will see how, by mixing transfer learning, zero-shot learning, and ONNX runtime, we can access the power of BERT right now, optimizing time and resources, achieving noticeable results on day one.

ghostwriter19 @ SardiStance: Generating new Tweets to Classify SardiStance EVALITA 2020 Political Tweets

Understanding the events and the dominant thought is of great help to convey the desired message to our potential audience, be it marketing or political propaganda.
Succeeding while the event is still ongoing is of vital importance to prepare alerts that require immediate action. A micro message platform like Twitter is the ideal place to be able to read a large amount of data linked to a theme and self categorized by its users using hashtags and mentions.
In this research, I will show how a simple translator can be used to bring styles, vocabulary, grammar, and other characteristics to a common factor that leads each of us to be unique in the way we express ourselves.

Ensemble of LSTMs for EVALITA 2018 Aspect-based Sentiment Analysis task (ABSITA)

In identifying the different emotions present in a review, it is necessary to distinguish the single entities present and the specific semantic relations. The number of reviews needed to have a complete dataset for every single possible option is not predictable. The approach described starts from the possibility to study the aspect and later the polarity and to create an ensemble of the two models to provide a better understanding of the dataset.

The validity of word vectors over the time for the EVALITA 2018 Emoji prediction task (ITAmoji)

This document describes the results of our system in the evaluation campaign on the prediction of Emoji in Italian, organized in the context of EVALITA 20181 (Ronzano et al., 2018). Given the text of a tweet in Italian, the task is to predict the emoji most likely associated with that tweet among the 25
emojis selected by the organizers. In this report, we describe the three proposed systems for evaluation.
The approach described starts from the possibility of creating two different models, one for the part of categorization, and the other for the part of polarity. And to combine the two models to get a better understanding of the dataset.