Dr. Dileep Kalathil received the NSF Faculty Early Career Development (CAREER) Award for his proposal titled “Towards a Principled Framework for Resilient, Data Efficient and Scalable Reinforcement Learning for Control.” He will use this award to address three significant challenges of the artificial intelligence evolution — resiliency, scalability and data efficiency of the system.
To tackle these challenges, he is using reinforcement learning principles and investigating the issue of scalability so that machine-learning systems can be integrated for large-scale technologies, such as massive power systems. He is also exploring ways in which progress in this area of machine learning can continue, despite the fact that there is limited data available.
“One thing I believe we should do as an engineering department is to give students the opportunity of experiential learning. They should be able to try things, work on real-world problems and act as engineers.”
Kalathil will also utilize an experiential learning approach to integrate this reinforcement learning research into his educational curriculum by working with students on the Aggie Deep Racer project — a tiny autonomous toy car used for testing reinforcement learning models. The idea is that the algorithm can be integrated into the toy car’s system and the application can be put to practice on a real track.
“One thing I believe we should do as an engineering department is to give students the opportunity of experiential learning,” Kalathil said. “They should be able to try things, work on real-world problems and act as engineers.”
Other domains of applicability for reinforcement learning include the smart grid infrastructure, autonomous driving, natural language processing, health care and gaming applications. Recently, there has been a lot of interest in using reinforcement learning in power systems applications, especially in the control of distributed energy resources. Electricity systems are undergoing a dramatic transformation with the increasing penetration of renewable generation and proliferation of distributed energy resources (DERs), such as electric vehicles, electricity storage, rooftop photovoltaic panels and smart heating, ventilation and air conditioning systems. The capacity of DERs in the U.S. is expected to reach around 400GW by 2025 with a total investment of $80 billion. Reinforcement learning can dramatically reduce the operating cost, increase the efficiency and significantly increase the reliability of the overall energy system.