Objective and Summary

My obsession is the development of tools and techniques for the interpretable, safe, and fair deployment of intelligent systems.

Machine Learning + Silicon Design Startup

My current development and research interests are in making ultra-low power neural network inference feasible through work at Syntiant. To date we have shipped more neural ASICS than any silicon provider in the world for problems ranging from voice interfaces to to sensor fusion. For a full professional history, please view my LinkedIn profile.

Social Impact Work

I spend my free time developing process for the IBM Watson AI XPRIZE as the technical lead for the prize. I represent the foundation in the Partnership on AI, which has included developing the AI Incident Database. Additionally, I previously chaired the Partnership's Safety Critical AI expert group with Ashish Kapoor and Amanda Askell.

Open Source Work

I am also very active in the development of open source code and organizations. I developed a federally-recognized 501c3, the Privly Foundation, dedicated to online privacy education. The foundation's activities included developing Open Source software, technology workshops, and supervising student developers. My current open source work focuses on the AI Incident Database.

PhD Research

My grad school career covered four distinct areas. First, I developed a simulator for fire, forest growth, timber sales, and weather. Second, I developed visual analytic tools for exploring the policy space of Markov Decision Processes (MDPs), including the wildfire simulator. Since many MDPs are defined by computationally expensive simulators, I next developed a surrogate modeling method that brings interactive specification of policy, reward, and optimization functions to large state space Markov Decision Processes. My final area of focus is Bayesian policy search using the surrogate model I developed.

Details on each of these areas are covered in specific cases below.


CV Items

  • Filters

      Filter CV Items by Type

      Filter CV Items by Topic

Chapters

First Year Results from the IBM Watson AI XPRIZE: Lessons for the "AI for Good" Movement

McGregor, S., Banifatemi, A.

This chapter summarizes the strengths and weaknesses of artificial intelligence in solving problem domains advancing the common wellfare. The chapter will be published with the competition and team details of the NIPS Competition Track.

I will provide the text upon request.

Citation  2018

Connecting Conservation Research and Implementation: Building a Wildfire Assistant

McGregor, S., Houtman, R., Metoyer, R., & Dietterich, T. G.

This chapter binds my PhD dissertation together with other works in AI applied to problems in conservation. If you want to learn more about my PhD work, I recommend starting here.

The attached version is pre-typeset editing from Cambridge Press. A preview of the book is available here.

Paper Citation BibTeX  2019

Introduction to NIPS 2017 Competition Track

(Many Authors)

This chapter surveys the machine learning competitions collected within the NIPS Competition Track. All chapter writers were invited to contribute to this survey chapter.

I will provide the text upon request.

Citation BibTeX  2018

Papers

MDPvis: An Interactive Visualization for Testing Markov Decision Processes

McGregor, S., Buckingham, H., Houtman, R., Montgomery, C., Metoyer, R., & Dietterich, T. G.

This paper is an extended abstract of the paper presented at the IEEE Symposium on Visual Languages and Human-Centric Computing.

Paper Presentation Slides Citation BibTeX  2015

Visualizing High-Dimensional MDPs with Model-Free Monte Carlo

McGregor, S., Houtman, R., Montgomery, C., Metoyer, R., & Dietterich, T. G.

Best Paper Nomination

This work extends Model-Free Monte Carlo to large state space Markov Decision Processes (MDPs) by factoring the state space. We successfully generate surrogate trajectories for a large state space MDP (more than 10^1000000 states) with 8 state features. We posted an extended version of the paper to arXiv.

Paper Citation BibTeX  2017

Facilitating Testing and Debugging of Markov Decision Processes with Interactive Visualization

McGregor, S., Buckingham, H., Dietterich, T. G., Houtman, R., Montgomery, C., & Metoyer, R.

This paper introduces Markov Decision Processes to visualization researchers as a theoretical formulation for sequential optimization problems. We explain the MDP formulation, present a visual analytic system exploring simulator-defined MDPs, and evaluate the system on a high dimensional MDP domain.

Paper Presentation Slides Citation BibTeX  2015

Interactive visualization for testing Markov Decision Processes: MDPVIS

McGregor, S., Buckingham, H., Dietterich, T. G., Houtman, R., Montgomery, C., & Metoyer, R.

The development of a computationally expensive MDP simulator demonstrated the absence of adequate tooling for sequential decision making problems. This paper presents a visual analytic system, MDPvis, that is easy to interface with MDP simulators and optimizers and extensible with additional visualizations.

Demo Paper Citation BibTeX  2017

Preventing Repeated Real World AI Failures by Cataloging Incidents: The AI Incident Database

McGregor, S.

Mature industrial sectors (e.g., aviation) collect their real world failures in incident databases to inform safety improvements. Intelligent systems currently cause real world harms without a collective memory of their failings. As a result, companies repeatedly make the same mistakes in the design, development, and deployment of intelligent systems. A collection of intelligent system failures experienced in the real world (i.e., incidents) is needed to ensure intelligent systems benefit people and society. The AI Incident Database is an incident collection initiated by an industrial/non-profit cooperative to enable AI incident avoidance and mitigation. The database supports a variety of research and development use cases with faceted and full text search on more than 1,000 incident reports archived to date.

The camera-ready paper is forthcoming. You can interact with the website today.

Demo Citation BibTeX  2021

FlareNet: A Deep Learning Framework for Solar Phenomena Prediction

McGregor, S., Dhuri, D., Berea, A., Muñoz-Jaramillo, A.

This workshop paper and poster introduce the work product of the 2017 NASA Frontier Development Lab (FDL) research team in space weather prediction. We modeled solar activity with deep neural networks to predict solar flares. FlareNet shows promise in predicting natural phenomena that can cause immense damage to human terrestrial infrastructure.

Paper Poster Citation BibTeX  2017

Allowing a Wildfire to Burn: Estimating the Effect on Future Fire Suppression Costs

Houtman, R. M., Montgomery, C. A., Gagnon, A. R., Calkin, D. E., Dietterich, T. G., McGregor, S., & Crowley, M.

This paper outlines the wildfire simulator I collaboratively developed for evaluating wildfire suppression decisions. Wildfires can reduce timber harvests in the present, but suppression decisions must balance firefighting costs, reduced fuel risks post-fire, and timber revenues.

Paper Citation BibTeX  2013

Factoring Exogenous State for Model-Free Monte Carlo

McGregor, S., Houtman, R., Montgomery, C., Metoyer, R., & Dietterich, T. G.

This work shows how to reduce the simulation time of a high-dimensional ecological Markov Decision Process (MDP) from 7 hours to less than a seconds. We apply the method outlined in the RLDM paper.

 2017

Toward Visualization Methods for Interactive Improvement of MDP Specifications

McGregor, S., Dietterich, T. G., & Metoyer, R.

This workshop paper presents an early version of MDPvis and discusses the need for stronger visualizations in the reinforcement learning community.

Paper Presentation Slides Citation BibTeX  2014

Fast Optimization of Wildfire Suppression Policies with SMAC

McGregor, S., Houtman, R., Montgomery, C., Metoyer, R., & Dietterich, T. G.

This paper is the first application of random forests for direct policy search. Additionally, it is the first case of optimizing wildfire suppression policies across 100 year time spans. This is a standalone version of the paper incorporated into the book chapter, "Connecting Conservation Research and Implementation: Building a Wildfire Assistant."

Paper Citation  2017

Fast Simulation for Computational Sustainability Sequential Decision Making Problems

McGregor, S., Houtman, R., Buckingham, H., Montgomery, C., Metoyer, R., & Dietterich, T. G.,

This paper, presentation, and poster give preliminary results for modeling the wildfire policy MDP with Model-Free Monte Carlo.

Video Paper Poster Presentation Slides Citation BibTeX  2016

Year One of the IBM Watson AI XPRIZE: Case Studies in "AI for Good"

McGregor, S., Banifatemi, A.

This is a summary of judging processes developed for the first milestone competition of the IBM Watson AI XPRIZE, as well as the teams that were nominated for the competition. I will provide the text upon request.

Citation BibTeX  2018

Presentations

AI for Good via Machine Learning Challenges

Banifatemi, A., McGregor, S., Chklovski, T., and Delisle, L.

In this invited talk I presented with three other people on our individual efforts at improving the world through incentivized AI competitions. I delivered my invited talk in two parts,

Part 1, IBM Watson AI XPRIZE: High Stakes "Academic Review": Here I detailed the collaborative ranking method I developed for the IBM Watson AI XPRIZE, which I will prepare for publication towards the end of 2020. The method was successful in efficiently selecting the strongest set of candidate teams for advancement in the IBM Watson AI XPRIZE.

Part 2, Inverting the Competition: This section was forward looking and explored approaches to competing over the production of useful data rather than useful models. This work is still in the early stages.

You can find more information on the workshop here and the abstract for our invited talk is below.

"AI for Good" efforts (e.g., applications work in sustainability, education, health, financial inclusion, etc.) have demonstrated the capacity to simultaneously advance intelligent system research and the greater good. Unfortunately, the majority of research that could find motivation in real-world "good" problems still center on problems with industrial or toy problem performance baselines. 
Competitions can serve as an important shaping reward for steering academia towards research that is simultaneously impactful on our state of knowledge and the state of the world.  This talk covers three aspects of AI for Good competitions. First, we survey current efforts within the AI for Good application space as a means of identifying current and future opportunities. Next we discuss how more qualitative notions of "Good" can be used as benchmarks in addition to more quantitative competition objective functions. Finally, we will provide notes on building coalitions of domain experts to develop and guide socially-impactful competitions in machine learning.
Citation  2019

Machine Learning in the Real World

McGregor, S.

This was an opportunity to present to a meetup group I founded several years ago in Portland Oregon. I am delighted to see the meetup group outlived my organizing and invited me back after more than 2 years away to present to the group.

Abstract: Companies and governments increasingly deploy intelligent systems to the real world. These systems--depending on the design and purpose--have the capacity to either greatly enhance the general well-being of the world or bring about some form of a science fiction dystopia.

I will cover both aspects of these systems in equal measure. First, I will share examples of leading-edge "AI for Good" research. Then, I will guide us through some real-world harms caused by deployed AI systems as explored from the AI Incident Database now under development with the Partnership on AI. Sean will then wrap up the talk with a discussion of current opportunities and risks in the development of AI systems for the general welfare.

Citation BibTeX  2019

Solving Grand Challenges with Artificial Intelligence

McGregor, S.

I keynoted the Thrival Innovation Festival on the topic of Artificial Intelligence for good. This talk was to a general business and startup audience and centered around the teams competing in the IBM Watson AI XPRIZE. Following the keynote I sat on a panel with David Danks (CMU Philosophy department chair), Amrit Dhir (Google for Entrepreneurs), Jana Eggers (Nara Logics), and Iba Masood (TARA.AI).

Presentation Slides Citation  2017

Making Your Privacy Software Usable

Davidson, J., & McGregor, S.

I collaborated with UX researcher Jen Davidson on a presentation of usability evaluation methods and privacy software ethics.

Citation BibTeX  2014

Digital Transformation and Collaboration in post-COVID world

Arun Rajasekaran, Christine Roberts, Sean McGregor, Sunil Madan

ValleyML is a volunteer-driven cross-industry community for advancing AI to empower people with a global reach of more than 200,000 professionals. In this electronic panel discussion, Arun Rajasekaran (Head, Technology Strategy, Wearables, Poly) led a discussion around electronic collaboration from the perspectives of parts makers (Syntiant), device makers (Poly), and service providers (Sunil Madan from Zoom). The panel was part of a larger convening taking place over the whole summer that covered many aspects of machine learning in industry. A video of the event is available on YouTube.

Panelists

  • Arun Rajasekaran: Head, Technology Strategy, Wearables, Poly
  • Christine Roberts: Senior Vice President and General Manager, Poly’s Enterprise Headset Division
  • Sean McGregor: ML Architect, Syntiant
  • Sunil Madan: Corporate CIO, Zoom
Video Citation  2020

Interdisciplinary Collaboration: Finding new Ways of Extracting Scientific Insights from Solar Observations.

McGregor, S., Munoz-Jaramillo, A., Dhuri, D., & Berea, A.

I spent the summer of 2017 working collaboratively with computer scientists and heliophysicists to predict solar flares from raw Solar Dynamics Observatory images. This presentation was an invited talk to present our preliminary results to the NASA community.

My contributions to this research included defining the software architecture for heliophysics deep learning research, including visualization methods for extracting scientific insights from trained neural network models.

Since the development process was unusual by NASA standards, the first quarter of the presentation focused on our style of agile development methods as a means of conducting interdisciplinary research.

Citation BibTeX  2017

Analog Computing for Deep Neural Networks

McGregor, S.

This talk thematically unifies the four design dimensions of power-efficient neural network computation, including, (1) changes to the neural network specification (e.g., depthwise separable convolutions), (2) changes to the numerical expression of the neural network (e.g., quantization), (3) post-training optimization of the neural network (e.g., pruning), and (4) custom silicon (e.g., analog matrix multiplication). This talk draws from my work at Syntiant developing neural network models for ultra-low power computation on analog silicon.

Citation  2018

Panel Discussion on Markets for Global Challenge

Alexandra Amouyel, Moustapha Cisse, Celina Lee, Irene Lo, Sean McGregor, Rudradeb Mitra, Victor Ohuruogu, Bright Simons, Milind Tambe, Nikhil Velpanur, Charity Wayua

A panel discussion broken into two segments. My fellow panelists included,

  • Alexandra Amouyel (MIT Solve)
  • Moustapha Cisse (Google; African Masters of Machine Intelligence)
  • Celina Lee (Zindi)
  • Irene Lo (Stanford University; Mechanism Design for Social Good)
  • Sean McGregor (XPRIZE Foundation and Syntiant)
  • Rudradeb Mitra (Omdena Inc.)
  • Victor Ohuruogu (UN Foundation Global Partnership for Sustainable Development Data)
  • Bright Simons (mPedigree)
  • Milind Tambe (Harvard University; Google)
  • Nikhil Velpanur (Wadhwani AI)
  • Charity Wayua (IBM Research Africa)
Citation  2019

Beneficial AI

United Nations, International Telecommunication Union

I served as panelist and/or rapporteur for this gathering of UN Agencies, NGOs, and academics. The group effort centered on developing guidlines and principles for the beneficial use of Artificial Intelligence. As part of my panel participation I drafted three documents in collaboration with panel moderators. These include one document on beneficial AI with the head of the Montreal Institute for Learning Algorithms (MILA), another document on privacy and security with the director of the United Nations Interregional Crime and Justice Research Institute, and a final document on measuring the positive and negative impacts of AI technologies with the head of UNICEF's venture fund.

Citation  2017

Impactful AI: Solving for Sustainable Development Goals

McGregor, S.

In this presentation I walked the UN World Food Program through the opportunity in applying AI to programs in food security and conservation. My goal was to simultaneously express the opportunities afforded by deployment of intelligent software products, while emphasizing that these solutions still require careful thought and clever formulations.

Presentation Slides Citation  2017

Moving Cloud Level Deep Learning Into Always On Battery Powered Edge Devices With Analog Neural Networks

McGregor, S.

This is a talk to the mems and sensors community, which is generally concerned with parts that go into devices. I presented within the emerging technologies track because our (Syntiant's) analog neural network is something both completely different and potentially transformative to the sensor community that produces more data than can conceivably be stored.

Citation  2019

AI for Good in Action

Banifatemi, A., McGregor, S., McConaghy, T.

This presentation surveyed recent results and organizational directions in the AI for Good space. My contribution to the presentation was to introduce a few examples of teams within the IBM Watson AI XPRIZE and to give the attendees an outline of the conference proceedings for which "AI for Good" is an explicit or implicit topic. I was later invited to reprise elements of the presentation at the Common Model Infrastructure Expo Workshop organized by Baidu.

Citation  2018

Incentives within Competitions for Social Good

McGregor, S.

The main idea of this workshop is that we should more deliberately think about the problem of facilitating interdisciplinary collaborations as a market design problem. We believe this is our research community’s path to greatest societal impact—we should focus on building the engine that allows people across disciplines to tackle important social issues, and in the process, expand the technical expertise of aspiring researchers in parts of the world where it’s currently lacking. Our goal is to take a project-based approach and provide apprenticeship learning to effectively mentor people into having novel intersections of domain and technical expertise.

My contribution to this topic is the attached video, which had live follow up questions. It covers

Part 1, IBM Watson AI XPRIZE: High Stakes "Academic Review": Here I detailed the collaborative ranking method I developed for the IBM Watson AI XPRIZE. The method was successful in efficiently selecting the strongest set of candidate teams for advancement in the IBM Watson AI XPRIZE.

Part 2, Inverting the Competition: This section was forward looking and explored approaches to competing over the production of useful data rather than useful models. This work is still in the early stages.

You can find more information on the workshop here and the complete video of the presentation is attached.

Video Citation  2020

Evaluating problems with AI

United Nations, International Telecommunication Union

In this talk I contextualized machine learning engineering processes for the mixed crowd of machine learning researchers and representatives of international organizations.

Citation  2019

Production Machine Learning Systems

McGregor, S.

Abstract: Over the last half century, the field of software engineering has produced tools and techniques for efficiently developing software. Some examples include unit/integration testing, integrated development environments, formal verification methods, and more. More recent machine-learned or "intelligent systems" pose distinctive challenges that are poorly served by these existing software engineering tools and techniques. This talk will cover the elements making intelligent system engineering distinctive from traditional software engineering and highlight tools and techniques tailored to the engineering challenges they introduce. This topic will be explored in the context of several motivating examples, including the development of a "production grade" speech recognition model at Syntiant, a system for making recommendations in response to wildfires, and collaborative work at the Partnership on AI.

 2019

How to Encrypt Your Content on Any Website

McGregor, S., Karve, S., & Davidson, J.

This presentation explained the technical foundations of the priv.ly project.

Citation BibTeX  2012

Panel Introducing the Facebook Deepfake Detection Challenge

Joaquin Quiñonero Candela, Cristian Canton Ferrer, Irina Kofman, Claire Leibowicz, Sean McGregor

As part of my service on the Partnership on AI's AI and Media Integrity Steering Committee, I sat on a panel for the launch of Facebook's Deepfake Detection Challenge. The panel focused on the importance of the deepfake detection problem and how we can mitigate the immediate impacts through the development of detector models. The central premise I presented to the panel was that we can build detector models for the current generation of deepfakes, but these detectors are brittle and will fail to detect with each generation of improved deepfake generator models. See my discussion of the topic here.

Citation  2019

Privacy (or not) in a Digital Age

McGregor, S.

This talk covered threats to electronic privacy and mitigation strategies. Topics covered included encryption, browser extensions, networking, and malware.

Citation BibTeX  2015

The Open Privacy Stack: Privly

McGregor, S., & Davidson, J.

OSCON is the premiere Open Source industry conference. This presentation focuses on our work on the priv.ly project's implementation.

Presentation Slides Citation BibTeX  2013

Posters

Incorporating Future Values into Analysis of Current Wildfires

McGregor, S., Houtman, R., Metoyer, R., Dietterich, T. G., Montgomery, C., & Crowley M.

Optimization and forestry are not often combined. We created this poster to gather input from forestry practitioners regarding the reports and visualizations that are useful in their work. The idea is to present our computer science research in a way that will lead to its adoption in a field that emphasizes field work over simulation students.

Poster Citation BibTeX  2014

Service

Mentor, Google Code In

I mentored students through the FOSSASIA mentoring organization in the 2016 Google Code In, a "...global, online contest introducing teenagers to the world of open source development" with variety of bite-sized tasks. My responsibility was to help students accomplish a variety of tasks advancing their skills and understanding of open source development. Mentoring for this program is both inspiring and tremendously valuable for developing cultural insights. Teenagers from around the world typically have not developed an international style of professional work so you see unfiltered communication and help develop a professional sense.

 2016

Organizer, Panel Moderator, Reviewer

I was an organizer for the NeurIPS Workshop focusing on social problems for which artificial intelligence has the potential to offer meaningful solutions. I additionally moderated the panel on "Academia, Corporations, Society, Responsibility", with Zoubin Ghahramani (Uber), David Danks (CMU), Lisa Di Jorio (Imagia), and Julien Cornebise (Element AI).

The workshop reach a peak attendance of more than 700 people during the lunchtime demonstration from acclaimed cellist Yo Yo Ma.

 2018

Co-Chair, Safety Critical AI Expert Group

The Safety Critical Expert Group brings together representatives from AI industrial organizations with persons from civil society organizations to address developing issues in AI safety. I serve as co-chair with Ashish Kapoor and Amanda Askell. My passion project within the working group is a database cataloging where intelligent systems have produced harms within the world.

 2019

General Organizer

I was an organizer for the NeurIPS Workshop focusing on social problems for which artificial intelligence has the potential to offer meaningful solutions. I worked to broaden the scope of the workshop from the prior iterations and include elements of producing good, characterizing bad, and policy for maximizing the former while minimizing the later.

 2019

Reviewer, NIPS

I reviewed reinforcement learning papers for this top-tier conference.

 2016

Steward, Coalition of Graduate Employees

The Coalition of Graduate Employees (CGE) is a labor union representing more than 1,600 graduate employees. During my time in graduate school I helped double the size of the bargaining unit and budget, bargain a multi-million dollar contract, collaborate with university human resources on the implementation of our health plan, serve as ranking member of our delegation to the state federation, and serve as secretary-treasurer.

 

Senior Program Committee Member, Session Chair

I served on the senior program committee for the "AI for Social Impact" track of the conference. My session largely concerned new and emerging datasets for social impact.

 2019-2020

General Organizer of NeurIPS Social

In 2019 the Neural Information Processing Systems conference added a peer-reviewed "socials track." The idea was to bring the social life of the conference back from sprawling corporate parties to gatherings organized by and for the research community. In this social event I submitted an abstract describing the event with my colleagues at the XPRIZE Foundation, then hosted a gathering attended to capacity.

 2019

Organizer, ICML Workshop on AI for Social Good

I was an organizer for the ICML Workshop focusing on social problems for which artificial intelligence has the potential to offer meaningful solutions. The organizing group shares responsibilities for chairing the review, schedule, and other event tasks.

 2019

Organizer, ICLR Workshop on AI for Social Good

I was an organizer for the ICLR Workshop focusing on social problems for which artificial intelligence has the potential to offer meaningful solutions. The organizing group shares responsibilities for chairing the review, schedule, and other event tasks.

You can find videos from the event on Slideshare.

 2019

Expert Group Member, Fairness, Transparency, Accountability Expert Group

The FTA brings together representatives from AI industrial organizations with persons from civil society organizations to address developing issues in AI fairness and accountability.

 2018

Founder, EECS Graduate Student Association

Despite having more than 400 graduate students in the department, I am the only one in my 6+ years to have organized events for the whole grad student population. I formed a student association and recruited a slate of officers to continue building on the social events I organized. I am told the organization has continued to hold events.

 2016

Founder, Privly Foundation

The Privly Foundation is a side project I kickstarted during my second year of graduate school. It is unrelated to my graduate studies, but has seen significant successes, including:

  • Designed and built method for posting private content to third-party social media: The Priv.ly Project
  • Gained attention in The Atlantic, Der Spiegel, Hacker News, and numerous other outlets.
  • Obtained 501(c)(3) tax exemption
  • Organization has mentored ten students in the Google Summer of Code (2013-2015)
  • Initiated online privacy meetup group in Portland, Oregon featuring presentations from leaders in security, incident response, citizen journalism, censorship circumvention, and cryptography
  • Funded at 260 percent of a Kickstarter campaign and raised total financial and in-kind support in excess of $100,000
 2011-2017