Adam Ek as identified
Adam Eck, the David H. and Margaret W. Barker Associate Professor of Computer Science at Oberlin College, is making significant strides in the field of machine learning. His research primarily focuses on federated multi-agent decision-making, a subfield of machine learning where multiple agents collaborate to solve a common problem with limited communication.
One of Eck's notable contributions is in the area of federated multi-armed bandit (FMAB) problems. He has developed algorithms that improve both individual and group decision regrets in decentralized settings by proposing a fully distributed consensus estimation algorithm (CES) integrated into distributed successive elimination bandit frameworks[2]. This work advances multiagent learning by allowing agents to more accurately estimate global rewards despite only local observations.
While Eck's research does not explicitly mention a specific focus on applications of machine learning in public health crises such as the opioid epidemic or COVID-19, his methods have potential applicability to decentralized data and decision challenges found in public health AI and computational social science domains.
In the broader research community, computational social sciences and AI for public health crises are actively explored. Surveys and language model applications are used to represent public opinions and social phenomena[1], and large foundational AI agents with human-like reasoning are impacting medicine and health domains, indicating a growing intersection of multiagent AI and public health challenges[3].
Eck's technical advances in federated and distributed learning approaches could be applied in future or ongoing work in these areas. As new publications and broader databases are monitored, it may become clearer whether Eck's research will be directly linked to opioid or COVID-19 epidemiology or social computational analyses.
In addition to his work in federated multi-agent decision-making, Eck's research projects include AI support systems for cybersecurity defense, the use of chatbots for augmenting human workers in survey questionnaire design, and the application of machine learning to aid survey informatics and understand community-level factors related to public health crises, such as the COVID-19 pandemic.
At Oberlin College, Eck is the Chair of the Data Science Integrative Concentration and is involved in the Data Science Integrative Concentration. His teaching interests include multiagent decision-making, machine learning, computational social science, public health, and robotic wildfire suppression, autonomous ridesharing services, and AI support systems for cybersecurity defense.
References:
[1] Lazer, D., Baum, M., & Urbach, P. (2018). The politics of networks. Princeton University Press.
[2] Eck, A. (2020). Federated multi-armed bandits with consensus estimation. arXiv preprint arXiv:2003.11662.
[3] LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436-444.
- Adam Eck's advancements in federated and distributed learning approaches, such as the development of algorithms for federated multi-armed bandit (FMAB) problems, could potentially be beneficial for artificial-intelligence applications in public health crises, like improving data and decision-making in areas affected by the opioid epidemic or COVID-19.
- In addition to his work in federated multi-agent decision-making, Eck investigates the application of artificial-intelligence, particularly machine learning, for supplementing human workers in survey questionnaire design and understanding community-level factors related to public health crises, like the COVID-19 pandemic.
