On the way to Self-learning and Knowledge Transmission across Generations

Considering the robot's potential in learning and handling uncertainties in unknown environment, will it be possible to allow a robot to explore and learn by itself?-- The self-learning discussed here, is a learning process that can be achieved without any user defined feature space--the feature space is rather defined and refined by the robot based on its goal and experience.

The purpose of studying self-learning is in the search of learning and passing knowledge across generations. Unlike many other creatures, a robot may be able to explore and transfer the knowledge across generations of other life beings. What information can a robot derive from the literature of our history? What can a robot learn from its own exploration and observation over the next few generations? Can robot sheds lights on our puzzles or help us to better understand ourselves and the universe?

Yet, despite the intelligence and potential of the robot that will be shown in the future development, there is no match between a robot and a real life. Each individual life is such a precious, unique entity that can reason, sense and express its feelings; it comes from the nature, and will eventually return to the nature; once life is over, there will be no repeat.

Creating robots, not only involves the techniques in making machines, but our soulfulness, and the respect for the life in all beings.

Contact

My current email address: dxiang at seas.upenn dot edu

Personal Website: http://homepages.cae.wisc.edu/~xdeng/simple/xiang.html