Theoretical Aspects of Machine Learning and Artificial Intelligence

PI计划在人工智能理论方面研究可证明有效的机器学习技术。要研究的一些主题是：通过探索学习未知系统或环境的技术，表示和利用先验知识的技术，交易假设复杂性以适应数据的技术，具有无限数量状态的确定性系统的学习以及具有概率转移或输出函数的有限状态概率系统的学习。

The PI plans to work on provably effective techniques for machine learning in the area of theoretical aspects of artificial intelligence. Some topics to be studied are: techniques for learning an unknown system or environment by exploration, techniques for representing and utilizing prior knowledge, techniques for trading hypothesis complexity for fit to the data, learning deterministic systems with an infinite number of states and learning finite-state probabilistic systems with probabilistic transitions or output functions.