Hi! My name is Dawson. I’m a second-year master’s student in the Energy and Resources Group (ERG) at UC Berkeley. My work focuses on the design and operation of modern energy systems. I’m especially interested in energy markets and forecasting. I take an interdisciplinary approach to these topics, drawing on tools from statistics, machine learning, economics, applied math, operations research, and public policy. You can reach me by email at verley [at] berkeley.edu.
Here are a few projects I’m working on:
Uncertainty Quantification for Modern Power Systems: Power system operators rely on estimates of uncertainty to inform operational decisions. In the Western Energy Imbalance Market, for instance, net load uncertainty intervals are used to determine flexible ramping requirements for each balancing area in the market. At the same time, uncertainty quantification is enjoying a renaissance in the statistical machine learning community, under the name “conformal prediction”. In ongoing work, I’m trying to understand how conformal prediction can support decision-making in energy systems, and also how conformal prediction compares to the simple heuristics currently used by energy systems engineers.
ML for Entity Resolution: Databases often contain multiple records that represent the same entity. In real-world applications, these records may not be linked by a unique identifier. In energy systems research, for instance, a single powerplant may be identified by name, address and operator in multiple EIA reports. Reliably merging these records to identify a single entity is essential for many modeling tasks. With collaborators at Stanford, I’m developing machine learning techniques for fast, large-scale entity resolution.
Other projects that are less well-developed are described here.
Here are some classes that inform my approach to research and teaching:
Undergraduate (at Stanford):
Graduate (at Berkeley):
At Berkeley, I was a Graduate Student Instructor (GSI) for the following courses: