[WiE Webinar] Autonomous decision-making in critical operations through complex Artificial Intelligence (AI) systems
Friday 4 April 2025 at 2:00 PM CET, 8:00 AM EST
By Rodica Mihai (NORCE Norwegian Research Centre, Norway)
Joint Webinar with IEEE Women in IES
Register now using the link below:
https://attendee.gotowebinar.com/register/2116671379466260830
Full Title: Autonomous decision-making in critical operations through complex Artificial Intelligence (AI) systems: key learnings from the research, development and demonstration in a real-world example
Abstract:
In critical operations, an AI system for decision-making should cope with poor observability, poor predictability and high-risk scenarios that can lead to highly damageable situations. At the same time domain knowledge plays a crucial role in the research and development of these solutions. These aspects will be touched upon in connection with a real-world case, from automation of drilling operations in geothermal and oil and gas.
The use case considered involves multiple providers and this is reflected on the data and information flow, adding a new level of complexity for any decision-making system. This raises several challenges including how to ensure stability of such a multi-agent system. When dealing with autonomous decision-making in critical operations, the management of transition from automatic/autonomous to manual mode is an important safety requirement and the algorithmic approach with domain constraints given by the use case will be presented.
The key learnings that will be addressed in the webinar include description of needs, challenges and current gaps in complex AI systems in critical operations based on the experience from the associated use case described.
Presenter’s bio:
Rodica Mihai is a senior researcher in the Energy modelling and automation group at NORCE Norwegian Research Centre. She has been working with autonomous decision-making systems in critical operations, from research to demonstration. Lately, she been focusing on verification and validation of complex AI systems and safe transition from autonomous to manual mode when such systems are in use. Mihai is currently leading the international SPE Drilling Systems Automation Technical Section subcommittee focusing on verification and validation of multi-agent automation systems with multiple providers. Mihai also holds a part-time position as associated professor at University of Bergen at Informatics Department, focusing on AI methods and systems.