Arianna Traviglia, the archaeologist 3.0 who teaches computers how to find finds –

“When I started studying archeology, I never imagined I’d end up analyzing data from satellites to discover ancient structures underground. And above all, I never thought I’d teach a machine how to identify them ». Arianna Traviglia is the director of the Center for Cultural Heritage Technology of the Italian Institute of Technology (Iit), based in Venice at the University of Ca ‘Foscari. Traviglia is a “returning brain”. After a PhD in geomatics and spatial information systems and a specialization in aerial and satellite image analysis, thanks to funding from the European Commission for the return of researchers to the EU, she returned to work in Italy after eight years at the University of Sydney and one in Seattle.

Traviglia was among the first in Italy to understand that archeology had to change its paradigm and open up to new tools. «Indiana Jones is a captivating character, but far from the everyday reality of the archaeologist. Going to the ground and digging will always remain central to our work, however there are new possibilities, ”explains Traviglia, according to whom future archaeologists will also be required to have computer skills and analysis of satellite photos. «I work in an interdisciplinary environment that collaborates with the European Space Agency through the Copernicus platform». Copernicus spreads the images of the Sentinel satellites, which cover the entire surface of the planet with their sensors every 5-6 days.

Teach the machines

To be able to automatically identify the buried archaeological deposits, it is first necessary to teach the machines to recognize them, a phase that is defined machine learning. Because even the most sophisticated program does not know what to do if it is not first taught how to move. “We have to teach machines to recognize structures. It is a long process because we have to transmit and replicate our mental processes of recognition on the machines », adds Traviglia. In this way, researchers will be able to see objects or irregularities that would otherwise be impossible for the human eye to see, such as traces in dense vegetation, or on bare terrain and hollows.

READ Also:  NASA intends to deliver a man to the moon earlier than Russia and China
Dataset in training

“We started a few months ago, there are few groups in this sector in the world. It’s a long job because, before starting the training of the machines, you have to build from scratch i dataset in training, i.e. the reference figures. Compared to the other groups, which focus only on the recognition of one or two types of trace, for example a round furnace or a rectangular enclosure, relatively simple shapes to teach, we instead want to teach to recognize any shape, for example a linear Roman road or a necropolis from the Bronze Age which has typical shapes. A machine only learns well if it is taught well ».

The project

The Cultural Landscapes Scanner Iit-Esa project by Dr. Traviglia is not only aimed at archeology, but could also be used in other fields. “If an area to be urbanized or where you want to build a shopping center hides archaeological structures that have not yet been unearthed, it is better to stop before starting, without wasting resources”. What are the models for training the computer system to recognize archaeological structures? «We use data collected in the area around Aquileia, then we will do other tests in Veneto, in Holland with the University of Leiden and on the Aran Islands in Ireland with the University of Glasgow, areas that have different archaeological structures from ours. I like to think that our pioneering studies will be used in the future for land control and to preserve cultural landscapes. These technologies allow us to preventively save what we have not yet discovered ».

READ Also:  The colonization of the moon by the principle of bees: Russian scientists have developed a new concept

April 6, 2021 (change April 6, 2021 | 21:38)

© Time.News


Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Recent News

Editor's Pick