Tutorials
The role of the tutorials is to provide a platform for a more intensive scientific exchange amongst researchers interested in a particular topic and as a meeting point for the community. Tutorials complement the depth-oriented technical sessions by providing participants with broad overviews of emerging fields. A tutorial can be scheduled for 1.5 or 3 hours.
Tutorial on
Intelligent control of IoT-based Systems
Instructor
|
Olivera Kotevska
Oak Ridge National Laboratory (ORNL)
United States
|
|
|
|
Brief Bio
Dr. Dipl-Ing. Olivera Kotevska is a research scientist in Computer Science and Mathematics Division at Oak Ridge National Laboratory (ORNL), in Tennessee, USA. She received her Ph.D. degree in Computer Science from the University of Grenoble Alpes, France. Before joining ORNL, she worked as an international guest researcher at the National Institute of Standards and Technology (NIST) in Washington, DC performing machine learning for smart city challenges as well as researching the security and privacy aspects of Home Internet of Things devices. Her doctoral research focused on using machine learning / artificial intelligence approach to systems based on Internet of Things to identify causal relations between events, sensors, and event prediction. Additionally, she is interested in differential privacy algorithms for edge computing systems and algorithms for increasing trust in AI-based systems. At ORNL, she works on using ML algorithms in event-based distributed/networked systems applied to effective sensing and control of complex systems such as electric grids, human mobility and biomedical.
|
Abstract
We live in a world of expansion of Internet of Things (IoT)-based systems. They are present in our daily life, work, manufacturing, transportation, industry and so forth. They help us to improve our life, processes, reduce energy, improve monitoring, and have better security. In order to do so we need to establish intelligent control of these networked IoT-based systems. Some of the newly developed machine learning (ML) algorithms such as reinforcement learning (RL) have this feature. RL also have capability to combine user preferences and history patterns when making the next decision. This tutorial will give overview of ML models with focus on RL with practical examples.
Keywords
IoT, control, machine learning, reinforcement learning
Aims and Learning Objectives
Participants will learn:
- the basics about IoT-based systems
- what intelligent control is and which algorithms can be used
- what is reinforcement learning and how to use it
Target Audience
Students, Professionals
Prerequisite Knowledge of Audience
Basic knowledge of machine learning. Would be beneficial if they tried at least one of the classification/regression algorithms.
Detailed Outline
- Internet of Things (IoT): Explanation (what they are and why they are useful), Applicability, Importance, Future of IoT
- Why we need to control them?
- The role of AI/ML in IoT control
- Taxonomy of AI/ML algorithms used for control.
- Reinforcement Learning (RL) as efficient approach for control
- Explanation of RL, types, focus on 3 types of RL algorithms (DQN, DDPG, multi-task DDPG), provide practical examples and results.
- Limitations of RL
- Conclusion