How AI can reveal new understandings of the past and the future

Holly Rushmeier, a computer science professor at Yale, is utilizing artificial intelligence (AI) to reconstruct the ancient city of Dura-Europos in Syria and assess land damage from Algerian forest fires. In the Dura-Europos project, Rushmeier s team is training networks to extract key contours from historic photos and gather facts about artifacts to create a knowledge graph for future research. This effort aims to consolidate scattered information from various sources into a linked open data system, such as Wikidata, to facilitate efficient question-answering. In a separate project, Rushmeier and her team are working with Nadia Zikiou, a Ph.D. student from Algeria, to analyze satellite data from European and U.S. satellites to assess damage caused by forest fires. The team is comparing convolutional neural networks (CNNs) and support vector machines (SVMs) to predict the extent of damage and monitor recovery. This work could help the Algerian government better manage their natural resources and plan for the future. Machine learning plays a crucial role in both projects. In the Dura-Europos project, it is used to train models to identify and categorize artifacts from images. In the forest fire project, machine learning is employed to analyze satellite data and predict the type and extent of damage caused by wildfires. The success of these models depends on having ample and appropriate training data, which can be used to improve the accuracy of future predictions and analyses.

Source: news.yale.edu
Published on 2024-09-11

Data Strategy launched by City to catapult service delivery into the digital age »

Open Data

Related news