From data
To meaning

We disclose the results of our research in two papers

In December, from 11th to 14th, IEEE BigData 2017 will take place in Boston, USA. This is a prestigious international conference about Big Data, that will host a workshop that will focus on the emergency management Data Science for Emergency Management (DSEM). We’re glad to participate in this conference with a paper written by Francesco Tarasconi, Michela Farina, Alessio Bosca e Antonio Mazzei, entitled: “The Role of Unstructured Data in Real-Time Disaster-Related Social Media Monitoring”. The event is supported by I-REACT (Improving Resilience to Emergencies through Advanced Cyber Technologies).I-REACT, a project coordinated by Istituto Superiore Mario Boella, aims to create a platform that integrate and analyses in real time all the useful data to manage emergencies for improving prediction and management of natural disasters. CELI is leader of two task for I-REACT project: “Linked Data and Semantic Structure” and “Social Media Data Engine”: we work on the semantic architecture of the system and on the analysis of social media (among which there are Facebook, Twitter and others, blogs and forums). The use of CELI’s technologies, among which Natural Language Processing, allows the survey and the comprehension of messages sent though the web related to natural disasters. This helps to integrate them in the platform. There are many workshops in this event, that aim to involve researchers, practitioners, environmental and governmental bodies to foster discussion on emergency management analytics open issues and provide interesting insights for future actions in the natural disaster management area.

The other paper is related to AgileRAI and it’s the result of a cooperation between RAI, Agile Lab and CELI. It’s entitled: “Real-time semantic enrichment of video streams in the age of Big Data”, written by Montagnuolo, Messina (RAI), Platter, Bidotti (Agile Lab) and A. Bosca (CELI). It has been presented in Big Data Tech, a conference that focuses on business and technology, the 12th of October in Milan. Furthermore, a more extended version of this paper has been accepted in the conference IEEE BigData 2017 that takes place in Boston, USA. In the project AgileRAI, it has been created a platform for searching, organizing and accessing multimedia data in a fast and semantic-driven way. We know that today television contents life cycle has a very long arch: they are produced, broadcasted and then a copy of them is stored in the archive to be reused or to be published online. In this project we worked on videos in an archive of contents; in order to be able to use this resource, that has a key role today for a media producer, it’s needed to enrich videos automatically, giving informations able to ease the research of them. In addition, given the growing amount of content produced everyday, the manual approach is not so clever, in the long term. The catalogue used for this project  has been enriched automatically with annotations related to different entities, for example monuments, paintings or places. AgileRAI is a system for the analysis of multimedia contents that faces the challenges by providing a scalable platform for innovative multimedia archiving and production services by leveraging advanced pattern recognition techniques and semantic annotation.

Tags: , , , ,