Title of the symposium on a blue circuitry background with EI and Sloan Foundation logo

Interested in this event? Recordings and presentations are available here.

How can machine learning and data science tools improve our understanding of energy systems and manage them to be more accessible, affordable, reliable, and clean?  This single-track symposium will explore cutting-edge approaches addressing this question, highlighting the work of established experts as well as emerging scholars in the field. 

In particular, the symposium will examine how data science and machine learning are driving innovation in areas including: 

  • Remote sensing of energy resources and infrastructure;
  • Energy systems modeling;
  • Energy consumption end uses;
  • Energy and climate change; and
  • Energy access. 

The Symposium will include a keynote presentation by Dr. Colin Parris (Senior Vice President and Chief Technology Officer, GE Digital); sessions featuring insights from industry and academic experts; breakout groups for networking; and quick talks by emerging scholars. The symposium is organized by the Energy Data Analytics Lab at Duke University and funding support is provided by a grant from the Alfred P. Sloan Foundation.

Due to abundant interest in this event, registration is now closed. Please subscribe to our newsletter at bit.ly/energyduke to be alerted when the event videos are published and to get email updates on future energy events and news at Duke.

  
Tuesday, December 8

12:30 - 12:50 Welcome

Opening Remarks

  • Dr. Brian Murray, Duke University
  • Dr. Evan Michelson, Sloan Foundation
12:50 - 1:35 Keynote Presenter

 View from Industry: Applying Artificial Intelligence to the Energy Sector & Beyond 

  • Dr. Colin Parris, GE Digital
1:35 - 1:45 Break  
1:45 - 2:55 Remote Sensing for Energy Resources and Infrastructure (Session 1)

Automatically Estimating Distributed Solar Photovoltaic Capacity and Building Energy Consumption Using Satellite Imagery 

  • Dr. Kyle Bradbury, Duke University 

Development Seed’s Work Using Machine Learning to Detect Energy Infrastructure 

  • Martha Morrissey, Development Seed 

Estimating GHG Emissions from Every Power Plant on the Planet 

  • Dr. Heather Couture, Pixel Scientia Labs 
2:55 - 3:05 Break  
3:05 - 3:45 Breakout Groups

Open networking and small-group topical discussions

3:45 - 4:55 Energy Systems Modeling (Session 2)

Surrogate Modeling for Scalable Assessment of Wind Plant Technology Innovation

  • Dr. Dylan Harrison-Atlas, National Renewable Energy Laboratory

Infrastructure Constraints on Energy Transitions: Lessons from History

  • Dr. Elisabeth Moyer, University of Chicago

Applying Data Science to Utility Operations

  • John Pressley and Dylan Lustig, Duke Energy Corporation
4:55 - 5:00 Day 1 Closing
  • Dr. Brian Murray, Duke University


Wednesday, December 9

12:30 - 12:40   Day 1 Recap
  • Dr. Kyle Bradbury, Duke University
12:40 - 2:10    Energy Demand (Session 3)

Harnessing Opportunistic Data and Physics Intuition to Enable Autonomous Buildings 

  • Dr. Mario Bergés, Carnegie Mellon University 

Counterfactual Prediction Methods as a Substitute for Randomized Control Trials: Evidence from High-Frequency Energy Data 

  • Dr. Brian Prest, Resources for the Future 

Decomposing 'the' Elasticity of Demand: Empirical and Policy Insights from 300 Million Natural Gas Bills 

  • Dr. Ed Rubin, University of Oregon 

CityDNN: A Deep Neural Network for Urban Energy Simulation Models 

  • Dr. Zoltan Nagy, University of Texas at Austin 
2:10 - 3:10 Breakout Groups Featured Lightning Talks: Announcing Research Video Competition awardees and small-group Q&A with lightning talk presenters
3:10 - 3:20  Break  
3:20 - 4:30 Energy Access and Reliability (Session 4)

Energy Access Explorer: A Mapping Platform to Connect SDG 7 and Sustainable Development Outcomes 

  • Dr. Dimitrios Mentis, World Resources Institute 

Recent work on Matching Methods for Causal Inference from Duke’s Almost-Matching-Exactly Lab 

  • Dr. Cynthia Rudin, Duke University 

Using Night Lights to Study the Politics of Energy Access and Reliability 

  • Dr. Brian Min, University of Michigan 
4:30 - 4:45 Closing Remarks Dr. Brian Murray, Duke University

All times PM Eastern

Dr. Colin Parris
 

Dr. Colin Parris, GE Digital
Senior Vice President and Chief Technology Officer
Dr. Colin Parris leads teams that work to leverage technologies and capabilities across GE to accelerate business impact and create scale advantage for digital transformation. Dr. Parris’ and his team’s work in artificial intelligence extend to the energy, aviation, and healthcare sectors. He created and leads the Digital Twin Initiative across GE. Dr. Parris has a Ph.D. in Electrical Engineering from the University of California, Berkeley.
Keynote Presentation: A View from Industry: Applying Artificial Intelligence to the Energy Sector & Beyond

 

Dr. Mario Berges
 

Dr. Mario Bergés, Carnegie Mellon University
Professor, Civil, and Environmental Engineering
Director, Intelligent Infrastructure Research Lab 

Dr. Mario Bergés is interested in making our built environment more operationally efficient and robust through the use of information and communication technologies, so that it can better deal with future resource constraints and a changing environment. His work largely focuses on developing approximate inference techniques to extract useful information from sensor data coming from civil infrastructure systems, with a particular focus on buildings and energy efficiency. Dr. Bergés has a Ph.D. in Civil and Environmental Engineering from CMU.
Presentation: Harnessing Opportunistic Data and Physics Intuition to Enable Autonomous Buildings

 

Dr. Kyle Bradbury
 

Dr. Kyle Bradbury, Duke University
Managing Director, Energy Data Analytics Lab 
Assistant Research Professor, Electrical and Computer Engineering 

Dr. Kyle Bradbury leads applied research projects at the intersection of machine learning techniques and energy problems. His work includes satellite imagery-based energy object identification, smart meter data analysis and disaggregation, and the cost/reliability tradeoffs of wind and solar integration into the grid through energy storage. He has a Ph.D. from Duke University focused on Energy Systems Modeling.
Presentation: Automatically Estimating Distributed Solar Photovoltaic Capacity and Building Energy Consumption Using Satellite Imagery

 

 

Dr. Heather Couture
 

Dr. Heather Couture, Pixel Scientia Labs
Founder, Pixel Scientia Labs  
Machine Learning Consultant & Researcher 

Dr. Heather Couture is the founder of the machine learning consulting firm Pixel Scientia Labs, which guides research and development teams to fight cancer and climate change more successfully with AI. Dr. Couture is currently working with WattTime on their Climate TRACE project for the electricity sector and has a Ph.D. in Computer Science from the University of North Carolina at Chapel Hill. 
Presentation: Estimating Greenhouse Gas Emissions from Every Power Plant on the Planet

 

Dr. Dylan Harrison-Atlas
 

Dr. Dylan Harrison-Atlas, National Renewable Energy Laboratory
Senior Data Scientist, Geospatial Data Science Group, Strategic Energy Analysis Center
Dr. Dylan Harrison-Atlas focuses on national scale assessment of energy problems and solutions using geospatial and machine learning methods. Active areas of research include investigating land use implications at the forefront of the energy-environment nexus and developing novel approaches to modeling technology innovation benefits across broad geographic extents. Dr. Harrison-Atlas has a Ph.D. in Ecology from Colorado State University. 
Presentation: Surrogate Modeling for Scalable Assessment of Wind Plant Technology Innovation

 

 

Dylan Lustig
 

Dylan Lustig, Duke Energy Corporation
Lead Analyst, Digital Transformation

Dylan Lustig works with Duke Energy’s Digital Transformation team. He works as a portfolio lead within Digital Transformation’s Information Management Solutions group and supports dozens of products and product teams across many domains of the company. With Duke Energy since 2016, he has also championed the adoption of new, Agile ways of working within IT, worked on the initial development of the company’s Internet of Things (IoT) platform, and worked in the regulatory and corporate finance space. 
Presentation: Applying Data Science to Utility Operations (with John Pressley)

 

 

Dr. Brian Min
 

Dr. Brian Min, University of Michigan 
Associate Professor, Political Science 
Faculty Affiliate for Energy Institute 

Dr. Brian Min is an Associate Professor of Political Science at the University of Michigan, Ann Arbor. Dr. Min studies the political economy of development with an emphasis on distributive politics, public goods provision, and energy politics. His research uses high-resolution satellite imagery to study the distribution of electricity, including collaboration with the World Bank to monitor electricity access in India and Africa. Dr. Min has a Ph.D. in Political Science from UCLA.
Presentation: Using Night Lights to Study the Politics of Energy Access and Reliability

 

 

Dr. Dimitris Mentis
 

Dr. Dimitris Mentis, World Resources Institute
Senior Energy Geographer, Energy Access
Project Lead, Energy Access Explorer 

Dr. Dimitris Mentis leads WRI’s energy access mapping efforts to prioritize areas where access to energy should be expanded while ensuring socio-economic development. He also founded and led the development of the Open Source Spatial Electrification Toolkit (OnSSET) in collaboration with multiple intergovernmental organizations, mapped the technical potential of renewable energy resources nationally and regionally through complex GIS assessments and focused on assessing the water energy nexus at different scales. Dr. Mentis has a Ph.D. in Energy and Environmental Systems from KTH Royal Institute of Technology, Sweden. 
Presentation: Energy Access Explorer: A Mapping Platform to Connect SDG 7 and Sustainable Development Outcomes

 

 

Martha Morrissey
 

Martha Morrissey, Development Seed
Machine Learning Engineer 
Martha applies machine learning models to detect urban infrastructure like uncommon building types and energy grids faster and more efficiently. Martha is passionate about bringing open geospatial data to academics and innovators as a basis for addressing large-scale global issues. Martha previously worked at Maxar Technologies (formerly DigitalGlobe) where she helped on API development for machine learning training data access, model creation, and model sharing. Martha has a Masters in Geography from the University of Colorado, Boulder.
Presentation: Development Seed's Work Using Machine Learning to Detect Energy Infrastructure

 

 

Dr. Elisabeth Moyer
 

Dr. Elisabeth Moyer, University of Chicago 
Associate Professor, Geophysical Sciences 
Director, Center for Robust Decision-making on Climate and Energy Policy (RDCEP) 

Dr. Moyer is an Associate Professor at the University of Chicago Department of the Geophysical Sciences and the director of the university's Center for Robust Decision-making on Climate and Energy Policy (RDCEP). Dr. Moyer’s research includes climate response to greenhouse-gas forcing; development of tools for impacts assessment; statistical emulation of climate model output; and climate and energy policy evaluation. Dr. Moyer has a Ph.D. in Planetary Science from the California Institute of Technology, and undergraduate degrees in Physics and Anthropology from Stanford University.
Presentation: Infrastructure Constraints on Energy Transitions: Lessons from History

 

 

Dr. Zoltan Nagy
 

Dr. Zoltan Nagy, University of Texas at Austin
Assistant Professor,  Civil, Architectural, and Environmental Engineering
Director, Intelligent Environments Laboratory

Dr. Zoltan Nagy is a roboticist turned building engineer, his research interests are in smart buildings and cities, renewable energy systems, control systems for zero emission building operation, machine learning and artificial intelligence for the built environment, complex fenestration systems and the influence of building occupants on energy performance. Dr. Nagy received a Ph.D. in Robotics from ETH Zurich, Switzerland.
Presentation: CityDNN: A Deep Neural Network for Urban Energy Simulation Models

 

John Pressley
 

John Pressley, Duke Energy Corporation
Director, Information Management Solutions, Digital Transformation

John Pressley works with Duke Energy’s Digital Transformation team, where he cut his teeth on scaling and supporting the Information Management Architecture (IMA). His portfolio includes over 32 products and platforms focused on pushing the company in new, innovative directions. He previously worked with Duke Energy’s HR IT team, as an analyst at Accenture in the High Tech and Telecomm division – with clients including Texas Instruments, Sony Erickson, Verizon, and Shell Oil, and Wachovia. John has an MBA from Wake Forest University. 
Presentation: Applying Data Science to Utility Operations (with Dylan Lustig)

 

Dr. Brian Prest
 

 

Dr. Brian Prest, Resources for the Future
Economist
Dr. Brian Prest uses economic theory and econometric models to improve energy and environmental policies by assessing their impacts on markets and pollution. This includes work on the impacts of federal tax credits for coal use, the social cost of carbon, econometric analysis of the oil and gas industry, analysis on the economic effects of rising temperatures, and household responses to time-varying electricity pricing. Dr. Prest has a Ph.D. in Environmental and Resource Economics from Duke University.
Presentation: Counterfactual Prediction Methods as a Substitute for Randomized Controlled Trials: Evidence from High-Frequency Energy Data

 

Dr. Edward Rubin
 

Dr. Edward Rubin, University of Oregon 
Assistant Professor, Economics 

Dr. Ed Rubin's research focuses on environmental/energy economics and inequality—particularly policy impacts, strategic responses to regulation, and measuring exposure/access. Ed is especially interested in combining causally informed econometrics/statistics with data science (machine learning, spatial analysis) to shed new light on previously resolved or unanswered questions. He received his Ph.D. in Agricultural and Resource Economics from the University of California, Berkeley.
Presentation: Decomposing 'the' Elasticity of Demand: Empirical and Policy Insights from 300 Million Natural Gas Bills

 

Dr. Cynthia Rudin
 

Dr. Cynthia Rudin, Duke University 
Professor, Computer Science 
Principal Investigator, Prediction Analysis Lab

Dr. Cynthia Rudin is a Professor of Computer Science, Electrical and Computer Engineering, and Statistical Science and directs the Prediction Analysis Lab, whose main focus is in interpretable machine learning, including applications to energy systems. She is also an Associate Director of the Statistical and Applied Mathematical Sciences Institute (SAMSI). Dr. Rudin has a Ph.D. in Applied and Computational Mathematics from Princeton University.
Presentation: Recent Work on Matching Methods for Causal Inference from Duke's Almost-Matching-Exactly Lab

 

The symposium will also highlight research by two interdisciplinary cohorts of doctoral students who are part of Duke University's Energy Data Analytics Ph.D. Student Fellows Program, funded by the Alfred P. Sloan Foundation

Five-minute presentations by emerging scholars (often Ph.D. students). Judges will award prizes to the top three videos, based on 1) Compelling communication of the core ideas and outcomes of the project in the video to an interdisciplinary audience; and 2) Innovation and potential for impact of the energy application and data science methodology. The competition will include the Lightning Talks listed below.

Alina Barnett

Ph.D. Student, Duke University

“Estimating Insulation Efficiency from Satellite Imagery using Super-Resolution”

Noman Bashir

Ph.D. Student, University of Massachusetts Amherst

“Solar-TK: A Data-driven Toolkit for Solar PV Performance Modeling and Forecasting”

Dipendra Bhattarai

Ph.D. Student, University of Tasmania, Australia

”Systematic Review on relation between Electricity Consumption and Night-time Lights”

Dr. Dong Chen

Faculty, Florida International University

“SolarFinder: Automatic Detection of Solar Photovoltaic Arrays”

Santiago Correa

Ph.D. Student, University of Massachusetts Amherst

“GridMapper: Mapping Electricity Grid Infrastructure in Developing Countries”

Priya Donti

Ph.D. Student, Carnegie Mellon University

“Inverse Optimal Power Flow: Assessing the Vulnerability of Grid Data”

Jan Drgona

Data, Scientist Pacific Northwest National Laboratory

“Combining Deep Learning and Physics for Cutting  the Costs of Advanced Building Control”

Juliana Felkner

Faculty, University of Texas at Austin

“Using Urban Energy Modeling to Assess the Impact of Building Retrofits on Energy Consumption and Overheating Risks in Future Climate Scenarios”

Dr. Todd Gerarden

Faculty, Cornell University

“Optimal Targeting of Information Provision Evidence from Home Energy Reports”

Bohao Huang

Ph.D. Student, Duke University

“Solar Array Mapping in San Diego”

Qingran Li

Ph.D. Student, Duke University

“Disaggregating the Behavioral Factors in Residential Load Profiles”

Sayak Mukherjee

Postdoctoral Fellow, Pacific Northwest National Laboratory

"Toward Data-Driven Control of Distributed Energy Resources using Reinforcement Learning" 

Olukunle Owolabi

Ph.D. Student, Tufts University

“Data-driven Application of Modern Portfolio Theory Risk Metrics of the Electrical Grid”

McKenna Peplinski

Ph.D. Student, University of Southern California

"Predicting Changes in Southern California’s Residential Electricity Consumption using Machine Learning Models”

Gopinath Rajendiran

Central Scientific Instruments Organisation, India

"Non-intrusive Load Monitoring (NILM) Technique Using Machine Learning Algorithms: Challenges and its Potential Applications towards Smart Sustainable Cities Development"

Jun Shepard

Ph.D. Student, Duke University

"Energy Flows through the Global Economy: 
From Production Source to Consumption Sink"

Ashwin Shirsat

Ph.D. Student, North Carolina State University

“Quantifying Residential Demand Response Potential Using Mixture Density Recurrent Neural Network”

Akansha Singh Bansal

Ph.D. Student, University of Massachusetts Amherst

“See the Light: Modeling Solar Performance using Multispectral Satellite Data” 

Edgar Virguez

Ph.D. Student, Duke University

“Utility-Scale Photovoltaics plus Storage: A Cost Effective Alternative for Decarbonization?”

Tianyu Wang

Ph.D. Student, Duke University

“Power Plant Pollution Monitoring via Bandit Algorithms”

Tongshu Zheng

Ph.D. Student, Duke University

“Estimating Solar PV Soiling Using a Satellite-Based Remote Sensing Approach”

Alfred P. Sloan Foundation LogoThe symposium is being organized by Dr. Kyle Bradbury, Dr. Jordan MalofDr. Brian MurrayDr. Billy Pizer, and Dr. Cynthia Rudin at Duke University's Energy Data Analytics Lab, a collaborative effort of the Duke University Energy Initiative (which houses it), the Rhodes Information Initiative at Duke, and the Social Science Research Institute. Funding support for the workshop is provided by a grant from the Alfred P. Sloan Foundation.

Note: Conclusions reached or positions taken by researchers or other grantees represent the views of the grantees themselves and not those of the Alfred P. Sloan Foundation or its trustees, officers, or staff.

Contact Trey Gowdy at trey.gowdy@duke.edu.

 

Location
Virtual Conference
Date & Time
Tuesday, Dec 08, 2020 - 12:30 pm to Wednesday, Dec 09, 2020 - 5:00 pm
Contact:
Jun
22
Location: Virtual
Time: 9:00 am to 4:00 pm