Adam Eck is an Assistant Professor in the Computer Science Department at Oberlin College. His primary research and teaching interests include: intelligent agents and multiagent systems, machine learning, data science, and computer-aided education.
More specifically, Adam enjoys learning about and developing solutions within decision making under uncertainty (how should agents gather information and behave to maximize rewards in complex, dynamic environments), reinforcement learning (how can agents learn how their worlds' operate in order to guide their decisions), and sequential supervised learning using recurrent neural networks (how can we predict future outcomes based on sequences of past observations).
Adam's projects include:
- reflective, deliberative information gathering using metareasoning about the benefits of information gathering to improve planning in uncertain environments (e.g., with applications to adaptive human-agent interactions)
- reinforcement learning to guide agent reasoning in open environments where agents appear and disappear over time, making collaboration and competition unstable and difficult to predict (e.g., with applications to wildfire suppression);
- supervised learning to predict outcomes of human user interactions with software, based on observed past behavior (e.g., with applications to predictions of breakoff and satisficing in online surveys)
- Survey Informatics, studying novel computational methods for improving the quality and quantity of data collection and analysis involving surveys and computer-assisted interviewing in the social sciences
- personal assistants for computer supported, collaborative learning