APRIL 26-27, 2022

Ben-Gurion University, in cooperation with the Israel Data Science Initiative (IDSI) and Helmholtz Information and Data Science Academy (HIDA), organized a two-day international hackathon.  The goal was to identify and map ancient water harvesting terraces in the vicinity of Evenari Avdat Farm, the Negev desert.  Groups of students from Germany, India, and Israel working in machine learning were supplied with large sets of aerial photos and terrain data, as well as computational resources.  The groups competed in designing deep learning models to predict the terrace locations based on “ground truth" data for training their models.

One of the mentors, Nadav Timor, posted:

“It was satisfying and rewarding to help teams of computer vision experts solve a semantic segmentation task and overcome barriers in utilizing deep learning for a real-life application. …. The wide variety of approaches, tools, and models was fascinating!…"

Initiation and preparation of the hackathon was a joint effort of Professor Arnon Karnieli and his team (Micha Silver, Arti Tiwari, and Aviv Cohen-Zada) of the Remote Sensing Lab, Sede Boker Campus, BGU, and Professor Mark Last and Rachel Raisa Eldin of the Data Science Research Center, BGU.  Academic, technical, and administrative support was provided by Professor (emeritus) Paul Feigin and Efrat Maimon of the IDSI, Technion, as well as colleagues from HIDA. 

The winning team members were Yi Wang and Chenying Liu, from the Data Science section of the German Aerospace Center (DLR)Their open-sourced code can be downloaded here.