Big Data, Artificial Intelligence and Machine Learning problems for Robust Food-Water-Energy NEXUS Analysis and Risk-Adjusted Agricultural Insurance Schemes

Ermolieva, T., Zagorodny, A., Bogdanov, V., Knopov, P., Gorbachuk, V., Zaslavski, V., Havlik, P. ORCID: https://orcid.org/0000-0001-5551-5085, Derci Augustynczik, A.L., Skalský, R. ORCID: https://orcid.org/0000-0002-0983-6897, Balkovič, J. ORCID: https://orcid.org/0000-0003-2955-4931, Folberth, C. ORCID: https://orcid.org/0000-0002-6738-5238, Kahil, T. ORCID: https://orcid.org/0000-0002-7812-5271, & Wang, G. (2022). Big Data, Artificial Intelligence and Machine Learning problems for Robust Food-Water-Energy NEXUS Analysis and Risk-Adjusted Agricultural Insurance Schemes. In: 11th International Conference “V.M. Glushkov’s Lectures”, 8th December 2022, Kiev, Ukraine.

Slideshow Presentation_Ermolieva_et_al_Glushkov's_leacture22_1312_22.pptx - Published Version Available under License Creative Commons Attribution Non-commercial. Download (3MB)

Abstract

In this talk we discuss the on-going joint work contributing to the IIASA-NASU 2022-2026 project on “Modeling and management of dynamic stochastic interdependent systems for food-water-energy-health security nexus” [1-3]. The research is being conducted involving the International Institute for Applied Systems Analysis (IIASA, Laxenburg, Austria), Committee for Systems Analysis (NASU, National Academy of Sciences, Ukraine), Glushkov’s Institute of Cybernetics (NASU, Kyiv, Ukraine)), Taras Shevchenko National University, Faculty of Computer Science and Cybernetics (Kyiv, Ukraine), Norwegian University of Science and Technology (Trondheim, Norway). Changing socio-economic and environmental conditions, increasing systemic interdependencies and risks, variability and frequency of weather-related disasters can have adverse impact on regional and global food security and, thereby, Food-Water-Energy (FEW) Nexus security. In this presentation we discuss the importance of appropriate Artificial Intelligence (AI), Statistical and Machine Learning (ML) model(s) for estimating and forecasting crop yield probability distributions relying on Big Data analysis, climate change projections, temperature and precipitation data, soil characteristics, land practices, and socio-economic indicators. Crop yield probability distribution models (scenario generators) can be integrated in agricultural insurance and land use models for the design of crop insurance mechanisms, which play a significant role in agricultural systems’ adaptation to climate change and systemic risks, as a part of the FEW systems’ adaptation strategies such as robust production technological portfolios, irrigation planning. The crop yield and agricultural production models can be linked with land use sectoral and regional models using models’ linkage methodologies based on distributed iterative “learning” optimization algorithms developed jointly by IIASA and NASU [4-8] and utilizing principles of Stochastic Quasigradient (SQG) methods of non-smooth stochastic optimization [4-7]. Novel ideas of systems’ linkage under asymmetric information and uncertainties enable the development of nested strategic-operational and local-global models. They are used in combination with Big Data, ML, AI (Deep Learning including deep neural learning or deep artificial neural network) and, in general, non-Bayesian probabilistic downscaling procedures for analyzing global and local systemic interdependencies and risks [1-4, 7-9]. Percentile-based indicators, goals and constraints are employed to analyze the robustness of estimates and policies to cope with systemic risks and extreme events.

Actions (login required)

View Item