|
|
|
- I lead the EVO-DesignOpt group.
- The challenges I am currently considering include:
Solutions for complex systems: from traffic to a desktop computer's activity, every system is becoming more complex. There's more data, less time and more things that interact. The problems are harder, bigger and more intertwined. Since nature solved this problem, I am considering how bio-inspired algorithms (on their own or in tandem with other techniques) can be practically deployed to accomodate the demands of dynamic adaptation or optimization. Current projects consider these challenges in the context of networks (for communication or biosynthesis via metabolic network knockouts), multi-core computer architecures and analog circuit CAD.
- Current projects in EVO-DesignOpt include
- Hierarchical genetic algorithms for efficient parallel sparse matrix algebra
- Genetic algorithms for network coding
- Genetic algorithms for biosynthesis
- Efficient, Scalable Evolutionary Algorithms for Multi-Objective Optimization Problems
- Hybrid search-based machine-learning methods capable of solving the general problem classes of classification and regression.
- Recent projects in EVO-DesignOpt include
- Evolutionary Algorithms for Analog Reconfigurable Systems
- Adaptive Resource Allocation for Multi-core processor environments
- If you are an MIT student, you are welcome to check EVO-DesignOpt or my current research pages for projects with possible UROP, UAP and M.Eng positions.
- I recently served as VP of Technology at Icosystem Corporation and led the Nymbler project.
- I no longer co-instruct Robotics: Science and Systems. Please refer your inquiries to the current teaching faculty.
- You may know me by my past membership in the Humanoid Robotics Group.
Copyright © MIT.
|
