Jordan Frank

Jordan Frank

Technion – Israel Institute of Technology

Haifa, Israel

Across the globe in Haifa, Israel, PhD student Jordan Frank undertook an internship at Technion – Israel Institute of Technology for a 2-month internship. Mr. Frank worked in close collaboration with Prof. Nahum Shimkin, a leading expert in both system identification and reinforcement learning.

Founded in 1924, Technion is the oldest university in Israel. The Department of Electrical Engineering at this world-class institution acts as a center of excellence in applied and theoretical research, contributing to the advancement of knowledge in electrical and computer engineering in Israel and throughout the world.

Mr. Frank’s work with Prof. Shimkin concentrated on dimensionality reduction and model construction for high-dimensional MDPs with applications to movement modeling. Most real-world problems have a high-dimensional, complex state space, while most of the existing approaches to sequential decision making require small, simple state spaces. In order to extend these approaches to tackle the real-world problems that they are designed to solve, one must find a way to reduce the dimensionality of the state space while maintaining its intrinsic properties. In particular, the research targeted human motion with the goal of modeling the motion as nonlinear dynamical systems on low dimensional manifolds from high-dimensional, noisy sensor data collected from wearable sensors: the task of creating a dynamical system from data is known as system identification. From these models, the group considered ways to perform planning tasks, using the principles of reinforcement learning.