The Colloquium on Digital Transformation is a series of weekly online talks on how artificial intelligence, machine learning, and big data can lead to scientific breakthroughs with large-scale societal benefit.

Fall 2020 Series

August 27, 2020, 1 pm PT/4 pm ET

Targeted Dynamic Interventions in Networked Epidemic Models

Asuman Ozdaglar, Distinguished Professor of Engineering Department Head, Electrical Engineering and Computer Science, Deputy Dean of Academics, Schwarzman College of Computing, MIT

Francesca Parise, Assistant Professor, Electrical and Computer Engineering, Cornell University

Epidemic spread models are playing an increasingly central role for policy making in the COVID-19 pandemic. Many of these models consider homogeneous populations, thus failing to capture rich heterogeneities in terms of risk factors, vulnerabilities, economic participation, location, and social interactions. In this colloquium, Asuman Ozdaglar and Francesca Parese present networked SIR models that capture groups of agents with different characteristics and interaction patterns. They then discuss targeted dynamic interventions for testing and lockdown policies that minimize spread of infection while containing social and economic damage. Their focus is on dynamic time-varying policies that adaptively adjust as a function of a community’s infection level.

Asu Ozdaglar

Asuman Ozdaglar is MIT’s MathWorks Professor of Electrical Engineering and Computer Science, Department Head of the Electrical Engineering and Computer Science Department, and Deputy Dean of Academics at the Schwarzman College of Computing. She is affiliated with Laboratory for Information & Decision Systems and the Operations Research Center. Her research focuses on problems that arise in the analysis and optimization of large-scale dynamic multi-agent networked systems including communication networks, transportation networks, and social and economic networks.

Francesca Parise

Francesca Parise was named Assistant Professor of Electrical and Computer Engineering at Cornell University in July 2020. She previously served as a postdoctoral researcher at the Laboratory for Information and Decision Systems at MIT. She was recognized as an EECS rising star in 2017, and is the recipient of the Guglielmo Marin Award from the Istituto Veneto di Scienze, Lettere ed Arti, the SNSF Early Postdoc Fellowship, the SNSF Advanced Postdoc Fellowship, and the ETH Medal for her doctoral work.

September 3, 2020, 1 pm PT/4 pm ET

Metapopulation and Age-Structured Epidemic Models for the COVID-19 Pandemic

Zoi Rapti, Associate Professor, Department of Mathematics, University of Illinois at Urbana-Champaign

In this colloquium, Zoi Rapti will first address the issue of parameter identifiability from reported data, such as confirmed cases and deaths. Parameter identifiability analysis is used to determine whether unknown parameters in ordinary differential equation models can be determined from the available data. She will then introduce age-structured and spatial (metapopulation) models for the spread of COVID-19, demonstrating that in some multi-group models that are structured by age, the basic reproductive number of each group is smaller than the basic reproductive number of the entire community. Hence, in most cases, it is not informative to consider separately groups that are known to interact, as this may underestimate the overall severity of the outbreak. In metapopulation models, she will show how traffic data is used in mobility models of epidemic spread. Rapti’s talk represents joint work of a research team that includes Eleftheria Kontou (Civil and Environmental Engineering, University of Illinois, Urbana-Champaign), Yannis Kevrekidis (Chemical and Biomolecular Engineering & Applied Mathematics and Statistics, Johns Hopkins University), and P. G. Kevrekidis (Mathematics and Statistics, University of Massachusetts at Amherst).


Zoi Rapti is an Associate Professor in the Department of Mathematics at the University of Illinois, Urbana-Champaign and the Carle R. Woese Institute for Genomic Biology. Her research focus is on nonlinear dynamics with applications to mathematical biology (DNA denaturation, epidemiology, community assembly, phage-bacteria interactions, data-theory coupling) and physics (Nonlinear Schrodinger-type equations, discrete Klein-Gordon equations).  She holds a bachelor’s degree in Mechanical Engineering from the National Technical University of Athens, Greece and a PhD in Mathematics from the University of Massachusetts, Amherst.

September 10, 2020, 1 pm PT/4 pm ET

Evolutionary Adaptations and Spreading Processes in Complex Networks

H. Vincent Poor, Michael Henry Strater University Professor of Electrical Engineering, Princeton University

A common theme among many models for spreading processes in networks is the assumption that the propagating object (e.g., a pathogen, in the context of infectious disease propagation, or a piece of information, in the context of information propagation) is transferred across network nodes without going through any modification. However, in real-life spreading processes, pathogens often evolve in response to changing environments or medical interventions, and information is often modified by individuals before being forwarded. In this talk, we will discuss the effects of such adaptations on spreading processes in complex networks with the aim of revealing their role in determining the threshold, probability, and final size of epidemics, and exploring the interplay between them and the structural properties of the network.


H. Vincent Poor is the Michael Henry Strater University Professor of Electrical Engineering at Princeton University, where he is engaged in research in information theory, machine learning and network science, and their applications in wireless networks, energy systems, and related fields, including recently modeling the spread of the COVID-19 epidemic. He is a member of the National Academy of Engineering and the National Academy of Sciences, and a foreign member of the Chinese Academy of Sciences and the Royal Society. Recognition of his work includes the 2017 IEEE Alexander Graham Medal and honorary doctorates from universities in Asia, Europe, and North America.

September 17, 2020, 1 pm PT/4 pm ET

Impact of Mobility on Epidemic Spread: Some Lessons from NYC and India

Saurabh Amin, Robert N. Noyce Career Development Associate Professor of Civil and Environmental Engineering, MIT

In this talk, we analyze the impact of mobility services (in particular, public transportation systems) on the spread of COVID-19. We also propose testing strategies based on mobility between areas with heterogeneous risk levels to control disease transmission. Our work focuses on two distinct regions: (1) NYC – where a significant number of people rely on the MTA for their daily commute; and (2) Odisha – an eastern state in India which saw a significant influx of migrant workers from COVID hotspots. In both regions, high levels of human mobility and limited testing capacity led to a rapid increase of COVID cases. In the study of NYC, we run panel analysis to evaluate the relative impact of MTA ridership over general mobility on the growth in cases at the zipcode level. We find strong heterogeneities across zip codes, which can be explained by socioeconomic factors and unbalanced testing resources. Importantly, we find that while a higher level of general mobility is associated with an increase in cases, the MTA usage did not lead to additional growth in cases after April. In contrast, the district-level data from Odisha exhibits strong correlation between case growth and volume of incoming migrant workers from high-risk states after May. We construct an optimization model that allows the state health authority to effectively allocate testing resources to worker populations based on the heterogeneous risk levels at their origin states and the local district population. This work is joint with Manxi Wu, Isabel Munoz, Devendra Shelar, and R. Gopalakrishnan.


Saurabh Amin is an Associate Professor in the Department of Civil and Environmental Engineering at MIT. He is affiliated with the Laboratory for Information and Decision Systems (LIDS) and the Operations Research Center at MIT. He received his PhD in Systems Engineering from UC Berkeley (2011), M.S.E. from UT Austin (2004), and B.Tech. from IIT Roorkee (2002). His fields of expertise include stochastic control theory, applied game theory, and optimization in networks. His research focuses on the design of high-confidence monitoring and control algorithms for infrastructure systems. He is a recipient of an NSF CAREER Award, the Google Faculty Research Award, the Robert N. Noyce Professorship, and the Ole Madsen Mentoring Award. 

September 24, 2020, 1 pm PT/4 pm ET

Towards AI for Healthcare with Applications to the COVID-19 Pandemic

Sanmi Koyejo, Assistant Professor, Department of Computer Science, University of Illinois at Urbana-Champaign

With the emergence of public healthcare crises such as the COVID-19 Pandemic, it is increasingly evident that AI is among the most powerful tools for addressing early detection of diseases, triage, treatment planning, and patient management, among many other pressing healthcare problems. What will it take to build effective machine learning systems for healthcare? This talk will outline our research progress towards answering this question. To this end, I will present emerging technical advances in federated learning, modeling, evaluation, privacy, and trustworthiness. I will also outline how we are bringing these tools to bear to aid in analyzing medical images from COVID-19 patients.


Sanmi Koyejo is an Assistant Professor in the Department of Computer Science at the University of Illinois at Urbana-Champaign. Koyejo’s research interests are in developing the principles and practice of adaptive and robust machine learning. Additionally, Koyejo focuses on applications to neuroscience and biomedical imaging. Koyejo has been the recipient of several awards including a best paper award from the conference on uncertainty in artificial intelligence (UAI), a Kavli Fellowship, an IJCAI early career spotlight, and a trainee award from the Organization for Human Brain Mapping (OHBM). Koyejo serves on the board of the Black in AI organization.

October 1, 2020, 1 pm PT/4 pm ET

Improving Fairness & Equity in Policy Applications of Machine Learning

Rayid Ghani, Distinguished Career Professor in the Machine Learning Department and the Heinz College of Information Systems and Public Policy, Carnegie Mellon University

We are in the early stages of using AI, ML, and Data Science to help make better policy decisions. Governments and nonprofits have started to explore how to improve society by tackling problems such as preventing children from getting lead poisoning, reducing police violence and misconduct, and increasing vaccination rates. As the impact of the COVID-19 pandemic has continued to increase, both ML researchers and practitioners have been proposing and developing methods to better understand, predict, and mitigate the spread, and governments have been exploring the use of these tools to better guide policy decisions. The AI and ML tools will hopefully be important elements of society’s efforts to overcome COVID-19, but as with any application of machine learning and AI, they also pose risks of introducing or exacerbating disparities. In this talk, I’ll give examples of recent work that highlights the use of ML/AI to achieve fair and equitable outcomes and challenges that need to be tackled in order to have social and policy impact in a fair and equitable manner.


Rayid Ghani is a Distinguished Career Professor in the Machine Learning Department and the Heinz College of Information Systems and Public Policy at Carnegie Mellon University. Rayid is a reformed computer scientist and wanna-be social scientist, focused on using large-scale Artificial Intelligence/Machine Learning/Data Science to solve public policy and social challenges in a fair and equitable manner. Rayid works with governments and non-profits on policy in health, criminal justice, education, public safety, economic development, and urban infrastructure. Rayid is passionate about teaching practical data science and started the Data Science for Social Good Fellowship that trains computer scientists, statisticians, and social scientists to work on data science problems with social impact. Before joining Carnegie Mellon University, Rayid was the Founding Director of the Center for Data Science & Public Policy, a Research Associate Professor in Computer Science, a Senior Fellow at the Harris School of Public Policy at the University of Chicago, and Chief Scientist of the Obama 2012 Election Campaign.

October 8, 2020, 1 pm PT/4 pm ET

Solving “Prediction Problems” in Health, from Heart Attacks to COVID-19

Ziad Obermeyer, Associate Professor of Health Policy and Management, University of California, Berkeley

In order to treat a disease, doctors must first know whether it is present. Thus, a key task in medicine is to judge the likelihood of a disease given rich, observable patient data. Because this resembles the prediction tasks where machine learning algorithms shine, we use them to study two important clinical problems. The first is the decision to test for heart attack. Because a test is only useful if it yields new information, efficient testing is grounded in accurate prediction of test outcomes. By comparing doctors’ testing decisions to tailored algorithmic predictions, we show that doctors both over-test (52.6% of high-cost tests for heart attack are wasted) and also under-test (many patients with predictably high risk go untested, then go on to experience frequent adverse cardiac events including death in the next 30 days). The second is the study of triage decisions in COVID-19. In emergency rooms across the world, doctors must decide if patients with suspected or confirmed disease are safe to go home, or if they need hospital-level monitoring. Clinicians have noted that many patients are tragically sent home, only to deteriorate rapidly. In ongoing work, we use machine vision to find subtle predictors of pulmonary collapse that human doctors miss. Together, these examples suggest that training algorithms to solve clinical “prediction problems” can yield both improvements in clinical care, and new insights into physician behavior and human health. 


Ziad Obermeyer is an Associate Professor at the University of California, Berkeley, where he does research at the intersection of machine learning, medicine, and health policy. He was named an Emerging Leader by the National Academy of Medicine and has received numerous awards including the Early Independence Award — the National Institutes of Health’s most prestigious award for exceptional junior scientists — and the Young Investigator Award from the Society for Academic Emergency Medicine. Previously, he was an Assistant Professor at Harvard Medical School. He continues to practice emergency medicine in underserved communities.

October 15, 2020 1 pm PT/4 pm ET

COVIDScholar: Applying Natural Language Processing at Scale to Accelerate COVID-19 Research

Gerbrand Ceder, Chancellor’s Professor, Department of Materials Science and Engineering, University of California, Berkeley

Amalie Trewartha, Postdoctoral Scholar, Division of Materials Science, Lawrence Berkeley National Laboratory

There is a critical need for tools that can help the COVID-19 researchers stay on top of the emerging literature and identify critical connections between ideas and observations that could lead to effective vaccines and therapies for COVID-19. To this end, our team at UC Berkeley and Lawrence Berkeley National Laboratory is building, a knowledge portal tailored specifically for COVID-19 research that leverages natural language processing (NLP) techniques to synthesize the information spread across more than 140,000 emergent research articles, patents, and clinical trials into actionable insights and new knowledge. Having its origins in our text-processing work in Materials Science, COVIDScholar is powered by an automated system that scrapes research documents from dozens of sources across the internet, cleans/repairs metadata as necessary, and analyzes the text with a number of NLP models for classification, information extraction, and scientific language modeling. We then integrate this information with specialized knowledge graphs which has the potential to give users unparalleled insight into the complex interactions that govern the transmission of COVID-19, the disease’s progression, and potential therapeutic strategies. This approach to combining textual information, such as word embeddings, with ontological knowledge graphs has the potential to improve the performance of machine learning models that operate on these data structures and to enable new ways of exploring literature on emerging subjects by leveraging past knowledge more efficiently.

Gerbrand Ceder

Gerbrand Ceder is the Chancellor’s Professor of Materials Science and Engineering at the University of California, Berkeley. His research is in computational and experimental materials design for clean energy technology and in Materials Genome approaches to materials design and synthesis. He has published over 400 scientific papers and holds more than 20 U.S. and foreign patents. He is a member of the U.S. National Academy of Engineering and the Royal Flemish Academy of Belgium for Science and The Art, a Fellow of the Materials Research Society and the Minerals, Metals & Materials Society, and has received awards from the Electrochemical Society, the Materials Research Society, the Minerals, Metals & Materials Society, and the International Battery Association. He is Co-Lead Scientist for new battery technologies at the U.S. Department of Energy’s Joint Center for Energy Storage (JCESR) and Chief Scientist of the Energy Frontier Research Center at the National Renewable Energy Laboratory (NREL).

Amelie Trewartha

Amelie Trewartha is a postdoctoral scholar in Gerbrand Ceder’s group at Lawrence Berkeley National Laboratory. She began her career as a nuclear physicist, before moving into materials science in 2019, with a focus on machine learning. Her research interests include the application of natural language processing (NLP) techniques to scientific literature, and building thermodynamically-motivated machine learning models for materials property prediction.

October 22, 2020, 1 pm PT/4 pm ET

Machine Learning-based Design of Proteins, Small Molecules, and Beyond

Jennifer Listgarten, Professor, Department of Electrical Engineering and Computer Sciences, University of California, Berkeley

Data-driven design is making headway into a number of application areas, including protein, small-molecule, and materials engineering. The design goal is to construct an object with desired properties, such as a protein therapeutic that binds tightly to its target. To that end, costly experimental measurements are being replaced with calls to a high-capacity regression model trained on labeled data, which can be leveraged in an in silico search for promising design candidates. The aim then is to discover designs that are better than the best design in the observed data. This goal puts machine learning-based design in a much more difficult spot than traditional applications of predictive modelling, since successful design requires, by definition, some degree of extrapolation–a pushing of the predictive models to its unknown limits, in parts of the design space that are a priori unknown. In this talk, I will anchor this overall problem in protein engineering and discuss our emerging approaches to tackle it.

Jennifer Listgarten

Jennifer Listgarten is a Professor in the Department of Electrical Engineering and Computer Sciences and the Center for Computational Biology at the University of California, Berkeley. She is also a member of the steering committee for the Berkeley AI Research (BAIR) Lab and a Chan Zuckerberg investigator. From 2007 to 2017, she was at Microsoft Research in Cambridge, MA (2014-2017), Los Angeles (2008-2014), and Redmond, WA (2007-2008). She completed her Ph.D. in the machine learning group in the Department of Computer Science at the University of Toronto, located in her hometown. She has two undergraduate degrees, one in Physics and one in Computer Science, from Queen’s University in Kingston, Ontario. Jennifer’s research interests are broadly at the intersection of machine learning, applied statistics, molecular biology, and science.

October 29, 2020, 1 pm PT/4 pm ET

Reliable Predictions? Counterfactual Predictions? Equitable Treatment? Some Recent Progress in Predictive Inference

Emmanuel Candès, Barnum-Simons Chair in Mathematics and Statistics, Professor of Statistics, and Professor, by courtesy, of Electrical Engineering, Stanford University

Recent progress in machine learning provides us with many potentially effective tools to learn from datasets of ever increasing sizes and make useful predictions. How do we know that these tools can be trusted in critical and high-sensitivity systems? If a learning algorithm predicts the GPA of a prospective college applicant, what guarantees do I have concerning the accuracy of this prediction? How do we know that it is not biased against certain groups of applicants? This talk introduces statistical ideas to ensure that the learned models satisfy some crucial properties, especially reliability and fairness (in the sense that the models need to apply to individuals in an equitable manner). To achieve these important objectives, we shall not “open up the black box” and try understanding its underpinnings. Rather, we discuss broad methodologies that can be wrapped around any black box to produce results that can be trusted and are equitable. We also show how our ideas can inform causal inference predictive. For instance, we will answer counterfactual predictive problems (i.e., predict what the outcome would have been if a patient had not been treated).

Emmanuel Candes, Dept. of Statistics

Emmanuel Candès is the Barnum-Simons Chair in Mathematics and Statistics, a Professor of Electrical Engineering (by courtesy), and a member of the Institute of Computational and Mathematical Engineering, all at Stanford University. Earlier, Candès was the Ronald and Maxine Linde Professor of Applied and Computational Mathematics at the California Institute of Technology. His research interests are in computational harmonic analysis, statistics, information theory, signal processing, and mathematical optimization with applications to the imaging sciences, scientific computing, and inverse problems. Candès has given over 60 plenary lectures in mathematics and statistics, biomedical imaging, and solid-state physics. Candès was awarded the Alan T. Waterman Award from the National Science Foundation and was elected to the National Academy of Sciences and the American Academy of Arts and Sciences in 2014.

November 5, 2020, 1 pm PT/4 pm ET

Reconstructing SARS-COV-2 Response Pathways

Ziv Bar-Joseph, FORE Systems Professor of Computer Science, Carnegie Mellon University

SARS-CoV-2 is known to primarily impact cells via two viral entry factors, ACE2 and TMPRSS2. However, much less is currently known about virus activity within cells. We used computational methods based on probabilistic graphical models to integrate several recent SARS-CoV-2 interaction and expression datasets with general protein-protein and protein-DNA interaction datasets. The reconstructed models display the pathways viral proteins use to drive expression in human cells and the pathways the cell uses to respond to the infection. Intersecting key proteins on these pathways with expression data from underlying conditions shown to increase mortality from SARS-CoV-2, and with knockout and phosphorylation data, identifies a few potential targets for treating cells to reduce viral loads.

Screen Shot 2020-10-30 at 7.28.39 PM

Ziv Bar-Joseph is the FORE Systems Professor of Computational Biology and Machine Learning at Carnegie Mellon University. His work focuses on the development of machine learning methods for the analysis, modeling, and visualization of time series high throughput biological data. Dr. Bar-Joseph is the recipient of the Overton Prize, an NSF CAREER Award, and several conference Best Paper awards. He is currently leading the Computational Tools Center for the National Institutes of Health Human BioMolecular Atlas Program (HuBMAP). He has served on the advisory board of several national efforts including the National Institute for Allergy and Infectious Diseases Systems Biology Program. Software tools developed by his group are widely used for the analysis of genomics data.

November 12, 2020, 1 pm PT/4 pm ET

Mathematics of Deep Learning

René Vidal, Herschel Seder Professor of Biomedical Engineering and Inaugural Director of the Mathematical Institute for Data Science, Johns Hopkins University

The past few years have seen a dramatic increase in the performance of recognition systems, thanks to the introduction of deep networks for representation learning. However, the mathematical reasons for this success remain elusive. For example, a key issue is that the neural network training problem is non-convex, hence optimization algorithms may not return a global minima. In addition, the regularization properties of algorithms such as dropout remain poorly understood. The first part of this talk will overview recent work on the theory of deep learning that aims to understand how to design the network architecture, how to regularize the network weights, and how to guarantee global optimality. The second part of this talk will present sufficient conditions to guarantee that local minima are globally optimal and that a local descent strategy can reach a global minima from any initialization. Such conditions apply to problems in matrix factorization, tensor factorization, and deep learning. The third part of this talk will present an analysis of the optimization and regularization properties of dropout for matrix factorization in the case of matrix factorization.

Rene Vidal

René Vidal is the Herschel Seder Professor of Biomedical Engineering and Director of the Mathematical Institute for Data Science at Johns Hopkins University. He is also an Amazon Scholar, Chief Scientist at NORCE, and Associate Editor in Chief of TPAMI. His current research focuses on the foundations of deep learning and its applications in computer vision and biomedical data science. He is an AIMBE Fellow, IEEE Fellow, IAPR Fellow, and Sloan Fellow, and has received numerous awards for his work, including the D’Alembert Faculty Award, J.K. Aggarwal Prize, ONR Young Investigator Award, NSF CAREER Award, and best paper awards in machine learning, computer vision, controls, and medical robotics.

November 19, 2020, 1 pm PT/4 pm ET

Tracking the Few and Far Between: Computational Strategies to Speed the Discovery of Low-frequency Genomic Variation in COVID-19

Nancy Amato, Head of the Department of Computer Science and Abel Bliss Professor of Engineering, University of Illinois at Urbana-Champaign

Lawrence Rauchwerger, University of Illinois at Urbana-Champaign

Todd J. Treangen, Rice University

To date, the vast majority of COVID-19 genomic research has been focused on a high-level view of SARS-CoV-2 diversity, overlooking the diversity of the viral population that exists within each COVID-19 positive patient. While viral load of SARS-CoV-2 in an individual can exceed hundreds of thousands of copies, available genomic databases only contain a single consensus version of this diverse population, discarding any low-frequency mutations. The goal of our project, CoVariants, is to develop novel computational approaches to recover these discarded variants and allow for rapid characterization of within-host diversity of SARS-CoV-2 across tens of thousands of samples. Commonly used computational tools either are not designed for the detection of low-frequency variants within viral populations, or require significant computational resources per sample. In this talk, we will describe how new parallelization strategies and approximate statistical methods can reduce the computational requirements of a widely used existing approach by up to 400 percent while preserving 100 percent of the low-frequency genomic diversity. We will end our talk by highlighting how we plan to use these improved computational methods to provide insight into the biological underpinnings of SARS-CoV-2 transmissibility and severity compared to other coronaviruses.


Nancy M. Amato is Abel Bliss Professor and Department Head of Computer Science at the University of Illinois. She received undergraduate degrees in Mathematical Sciences and Economics from Stanford, and M.S. and Ph.D. degrees in Computer Science from UC Berkeley and the University of Illinois, respectively. Her research focuses on robotics motion planning, computational biology and geometry, and parallel computing. Amato received the 2019 IEEE RAS Leadership Award in Robotics and Automation, the 2014 CRA Habermann Award, and the inaugural NCWIT Harrold/Notkin Research and Graduate Mentoring Award in 2014. She is a Fellow of the AAAI, AAAS, ACM, and IEEE.


Lawrence Rauchwerger is a professor in the Department of Computer Science at the University of Illinois. Previously, he was the Eppright Professor of Computer Science and Engineering at Texas A&M University and co-Director of the Parasol Lab. He received an engineering degree from the Polytechnic Institute Bucharest, a M.S. in Electrical Engineering from Stanford University and a Ph.D. in Computer Science from the University of Illinois. His approach to auto-parallelization, thread-level speculation and parallel code development has influenced industrial products at corporations including IBM, Intel, and Sun. Rauchwerger is a Fellow of the AAAS and IEEE and has received an NSF CAREER Award, awards from IBM and Intel.

ToddTreangen 500x500

Todd J. Treangen, Ph.D. is an Assistant Professor in the Department of Computer Science at Rice University and co-lead of the COVID-19 International Research Team. Before joining Rice, Dr. Treangen was a Research Scientist at the University of Maryland College Park. He received his Ph.D. in Computer Science in 2008 from the Polytechnic University of Catalonia (Barcelona, Spain). His research group focuses on solving large-scale computational problems specific to computational biology, with a focus on developing robust software tools targeted towards biothreat screening, infectious disease monitoring, and microbial forensics.

December 3, 2020, 1 pm PT/4 pm ET

Stochastic Optimization of Inventory at Large-scale Supply Chains

Mehdi Maasoumy, Principal Data Scientist,

Companies managing physical goods face growing supply chain challenges, given accelerated business cycles, stochasticity in supply and demand, and complex, interdependent, global supply chains. Enterprises often use inventory buffers – for raw materials, work-in-process goods, and finished products – as a way to help them manage service levels, and on-time and in full (OTIF) metrics. However, holding excess inventory often complicates business operations, and can lock up billions of dollars of capital that could be otherwise deployed.

Most manufacturing companies have historically deployed advanced Manufacturing Resource Planning (MRP) software, however, most commercially available MRP solutions do not result in optimal solutions for modern enterprises.

The C3 AI Inventory Optimization application fundamentally re-formulates and solves the supply chain and inventory problem as a constrained stochastic optimization problem. Our goal is to find the optimal re-order parameters that minimize inventory levels subject to a pre-defined service-level constraint and other operational constraints. The output of our approach can be the optimal placement of orders or the optimal inputs into an MRP system to ensure that the MRP heuristic then results in the right material levels and production operations.

Using C3 AI Inventory Optimization, we have helped our customers achieve inventory reductions of 30-50 percent‚ saving hundreds of millions of dollars at major enterprises at a global scale.

Mehdi Maasoumy
Mehdi Maasoumy is a Principal Data Scientist at, where he is leading AI teams that develop machine learning, deep learning, and optimization algorithms to solve previously unsolvable business problems — including Stochastic Inventory Optimization of Supply Chain and Predictive Maintenance across a wide range of industries such as Oil and Gas, Manufacturing, Energy, and Healthcare. Mehdi holds a M.Sc. and Ph.D. from the University of California at Berkeley and a B.Sc. from Sharif University of Technology, and is the recipient of three best paper awards from ACM and IEEE.

December 10, 2020, 1 pm PT/4 pm ET

Housing Precarity, Eviction, and Inequality in the Wake of COVID-19

Karen Chapple, Professor and Chair of City & Regional Planning, University of California, Berkeley

Tim Thomas, Research Director, Urban Displacement Project, University of California, Berkeley

Peter Hepburn, Assistant Professor of Sociology, Rutgers University-Newark

COVID-19 has the potential to exacerbate a severe housing and economic crisis in the United States, which will in turn undercut public health responses to the pandemic. Ensuring housing security is vital to mitigating the spread of the virus and sustaining health, economic security, and family stability. This joint, interdisciplinary project between the University of California, Berkeley and Princeton University brings together a group of academics and data scientists to track, analyze, and respond to pandemic-driven spikes in eviction and displacement risks. Doing so requires two central elements, both of which rely heavily on data science tools and methodologies. First, we have developed the Eviction Tracking System, an innovative tool for tracking real-time eviction filings in more than a dozen cities across the U.S. Second, we have developed a housing precarity risk model using machine learning that allows us to better analyze and predict areas at disproportionate risk of eviction, displacement, unemployment, and infection in the wake of the COVID-19 pandemic. This project provides major new sources of data that serve to inform research and public policy regarding housing and inequality in America.


Karen Chapple, Ph.D., is a city planner by training who studies inequalities in the planning, development, and governance of regions in the U.S. and Latin America, with a focus on economic development and housing. Her most recent book is Transit-Oriented Displacement or Community Dividends? Understanding the Effects of Smarter Growth on Communities.

Version 2

Tim Thomas is an urban sociologist, demographer, and data scientist. His research examines how neighborhood change impacts racial and gender disparities in housing, segregation, and forced mobility. His work has been published in academic journals and used as evidence to inform civil and state housing law.


Peter Hepburn is a sociologist and demographer. His research examines how changes to three core social institutions — work, criminal justice, and housing — serve to produce and perpetuate inequality. His work has been published in Social Forces, Demography, Social Problems, and the Journal of Marriage and Family.