DARPA Grant for Theodoros Rekatsinas
Together with colleagues from the University of Wisconsin-Madison, Professor Theodoros Rekatsinas has been awarded a four-year programme from the Defense Advanced Research Projects Agency (DARPA) for the joint project “LAKE”. Congratulations!
Professor Theo Rekatsinas, from the Institute for Computing Platforms at the Department of Computer Science, together with colleagues from the University of Wisconsin-Madison, has been granted a four-year programme from the Defense Advanced Research Projects Agency (DARPA) for the project “Large-scale Application of Neural Structured Learning to Accelerate Knowledge Extraction and Discovery (LAKE)”. The project, which started as part of DARPA's ASKE AI Exploration, will develop AI technology to streamline the discovery and extraction of knowledge from scientific publications.
Theodoros Rekatsinas is part of the Systems Group at ETH Zurich. His mission is to build a next-generation AI-powered technical assistant that extracts and assimilates data from heterogeneous sources to accelerate the discovery and extraction of scientific knowledge for subject matter experts. Scientific knowledge is encapsulated in ever-growing repositories of publications and computer codes that model the behavior of physical, biological, and social systems. Due to the heterogeneity and rate of production of the artifacts that populate these repositories, building and calibrating scientific models that facilitate agile response to emergent challenges requires automating, in near real-time, the identification, extraction, and aggregation of information from within publications and code bases. The project is part of the broader ASKEM DARPA programme, which aims to create a knowledge-modeling-simulation ecosystem to accelerate the development of scientific models with a focus on epidemics and space weather.