My research interests center on developing and using machine learning and data mining methods for real-world applications with a special interest in severe weather and in space. I am very interested in issues of representation and much of my current research focuses on developing spatiotemporal relational data mining methods and applying these to multiple real-world applications. Other interests include:
- Autonomous discovery of structure. Humans are exceptionally good at recognizing patterns in their interactions with the world and in using these patterns to successfully carry out a wide variety of tasks. Can we create methods that enable a computational agent to autonmously discover the salient structure of a task as it interacts with the real-world? Can the agent then use this structure to generalize and enable it to be successful at a wide variety of tasks, without having to be separately trained on each task?
- My dissertation
introduced a method that enabled an agent to autonomously identify and create useful temporal
abstractions from the agent's interaction with its environment.
- Current work focuses on identifying structure in using Bayesian, relational learning, and reinforcement learning.
- Knowledge representation. Humans generalize their knowledge from one task to related tasks yet robots and agents often have to be separately trained for each possible change in the environment. How can we efficiently
represent the knowledge learned in one task and reuse it for other
tasks? This knowledge can take the form of a control policy
learned to solve one task or a representation of structure in an
- Current work focuses on building task-oriented knowledge representations using relational learning methods
- Interaction and ensembles of learning methods. No single learning paradigm will work for every situation. I am interested in how the different paradigms can work synergisticall with one another to create a more robust autonomous system.
- Current work focuses on relational learning, supervised learning, Bayesian learning, reinforcement learning, and ensemble techniques.
- Improving STEM education and the retention and recruitment of underrepresented groups. To create a top science, technology, engineering, and mathmatics (STEM) workforce, we need a diverse and flexible workforce. Diversity will bring new ideas to the forefront and flexibility is required when technology is changing so rapdily. I'm interested in developing and applying innovations in STEM education that enable us to better train the scientists and engineers of tomorrow. I'm also interested in ways to significantly improve the diversity of our STEM workforce.
- Current work includes applying methods from top researchers in engineering education, mentoring the OU chapter of ACM-W (Association for Computing Machinery - Women's Chapter), and mentoring the OU chapter of Alpha Sigma Kappa (sorority for women in technical studies)
I direct the Interaction, Discovery, Exploration and Adaptation (IDEA) lab. You can find more information on our research from the lab web page.
My other pages
[Publications] [Courses] [I run a blog for scouting resources (focused on Boy Scouts of America)]
My Husband: Prof. Andrew Fagg
amy [at] cs.ou.edu
February 9, 2012 10:12 AM