Project description DEENESFRITPL The first automated causal discovery software engine Causal discovery methods are not new. They have been used to identify potential cause-effect relationships from observational data for decades. Even though they are becoming increasingly important in science and industries, they are largely inaccessible to non-experts. The EU-funded AUTOCD project proposes to create the first automated causal discovery software engine and explore its commercial exploitation. AutoCD will increase the productivity of experts and allow the application of causal discovery with minimal expertise. Such automation relies on the outcomes the CAUSALPATH ERC project. AutoCD will compare the development of the growing industry of automated machine learning (AutoML) libraries and platforms. Show the project objective Hide the project objective Objective Causal Discovery is desperately needed in both science and the industry, but it is largely inaccessible to non-experts. AutoCD proposes to create the first automated causal discovery software engine and explore its commercial exploitation. AutoCD will largely boost the productivity of experts as well as allow the application of causal discovery with minimal expertise. It will provide functionalities such as (a) induction of causal models and causal relations from data by automatically tuning the algorithmic causal discovery choices and their hyper-parameters, (b) inferences regarding the strength of causal effects and exploration of what-if scenarios of possible interventions. Such automation has only become recently possible due to research performed of the origin ERC named CAUSALPATH. We will work with two industrial partners, namely Gnosis Data Analysis and Huawei to validate AutoCD on real data and problems. Gnosis commercializes the JADBio product, which is a SaaS AutoML platform with obvious synergies to AutoCD. It has an expressed interest in AutoCD for a potential licensing deal (see letter of intent). AutoCD parallels the development of automated machine learning (AutoML) libraries and platforms that is growing to a $14bil industry. The project will create an MVP at TRL 5 and a business plan to commercialize the product. The research team (2 Profs, 1 Ph.D. student, 1 scientific programmer) have extensive collective experience not only inventing and designing novel causal discovery algorithms. In addition, the PI is also the co-founder of Gnosis with extensive experience in creating deep tech AutoML products and commercializing them. He will devote 70% of his research time to the project. Fields of science natural sciencescomputer and information sciencesdata sciencenatural sciencescomputer and information sciencessoftwaresocial sciencessociologyindustrial relationsautomationsocial scienceseconomics and businesseconomicsproduction economicsproductivitynatural sciencescomputer and information sciencesartificial intelligencemachine learning Programme(s) HORIZON.1.1 - European Research Council (ERC) Main Programme Topic(s) ERC-2022-POC1 - ERC PROOF OF CONCEPT GRANTS1 Call for proposal ERC-2022-POC1 See other projects for this call Funding Scheme ERC-POC - Proof of Concept Grant Coordinator PANEPISTIMIO KRITIS Net EU contribution € 150 000,00 Address University campus gallos 74100 Rethimno Greece See on map Region Νησιά Αιγαίου Κρήτη Ρέθυμνο Activity type Higher or Secondary Education Establishments Links Contact the organisation Opens in new window Website Opens in new window Participation in EU R&I programmes Opens in new window HORIZON collaboration network Opens in new window Other funding € 0,00
