Abstract: Women constitute approximately 50% of the population and have been an active part of the US workforce for over half a century. Yet women continue to be poorly represented in leadership positions within business, government, medical, and academic hierarchies. As of 2023, only 10% of Fortune 500 chief executive officers are women, 28% of US congressmembers are women, and 37% of practicing physicians are women. The decreasing representation of women at increasing levels of power within hierarchical professions has been called the “leaky pipeline” effect, but the main cause of this phenomenon remains contentious. Using a mathematical model of gender dynamics within professional hierarchies and a new database of gender fractionation over time, we quantify the impact of the two major decision-makers in the ascension of people through hierarchies: those applying for promotion, and those who grant promotion. The model is the first to demonstrate that intervention may be required to reach gender parity in some fields.

Bio:  Dr. Sara Clifton earned her Ph.D. at Northwestern University in Engineering Sciences and Applied Mathematics and then was a J.L. Doob Research Assistant Professor at the University of Illinois at Urbana-Champaign. She taught, mentored undergraduate research, and served as the director of the mathematical biology program at St. Olaf College, before joining Kenyon College in 2023. Her current research interests include biological dynamical systems and mathematical modeling of complex social systems. 

Abstract: Recent years have seen an explosion of scientific evidence showing that there is a different way to learn, lead and live, available to us all. When people take a limitless approach to learning – in learning and in life – different pathways open up, leading to higher, more equitable and more enjoyable achievement. In this session we will consider what this different approach is, thinking about the ways we can teach students mathematics to increase equity, engagement, and achievement.  We will see together the beauty of mathematics, across the grades, and learn how to bring that beauty to all students lives.

Bio: Dr Jo Boaler is the Nomellini & Olivier Professor of Education at Stanford University. Former roles have included being the Marie Curie Professor of Mathematics Education in England, and a maths teacher in London comprehensive schools. Her PhD won the national award for educational research in the UK. She is an elected fellow of the Royal Society of Arts (Great Britain), and a former president of the International Organization for Women and Mathematics Education (IOWME). She is the recipient of a National Science Foundation ‘Early Career Award’, the NCSM Kay Gilliland Equity Award (2014) and the CMC Walter Denham Mathematics Leadership award (2015). She is the author of eighteeen books and numerous research articles. She is a White House presenter on women and girls. She co-founded www.youcubed.org to give teachers, parents and students the resources and ideas they need to inspire and excite students about mathematics. Her work has been published in the New York Times, TIME magazine, The Telegraph, The Atlantic, The Wall Street Journal and many other news outlets. Her latest book is: Limitless Mind: Learn, Lead & Live without Barriers, published by Harper Collins. She is one of the writing team for the proposed Mathematics Framework for the state of California, co-leading a K-12 Data Science Initiative and was named as one of the 8 educators “changing the face of education” by the BBC.

Abstract: Over the past five years, it has been shown that idealized mathematical models of transactions between individuals can be used to accurately explain the very unequal distribution of wealth observed in free-market societies.  In stark contrast to neoclassical economics, these models indicate that free markets are naturally unstable, and that it is only exogenous redistribution that stabilizes them.  It is telling that models of exchange that display a natural tendency for wealth to “trickle upward” are able to explain empirical wealth distributions much more accurately than those that posit a supply-demand equilibrium.  Such models, called asset-exchange models, shed new light on the mechanisms underlying wealth accumulation, the growth of inequality, the dimnishing of upward mobility, and the origin and nature of oligarchy.  In so doing, they provide a new mathematical perspective to fields as diverse as economics, ethics, political science and public policy.  In this talk, we describe the construction and use of such mathematical models, comparing and contrasting the insights they provide with those of more orthodox economic theories.

Bio: Dr. Bruce M. Boghosian has been a Professor of Mathematics at Tufts University since 2000, and served as chair of Mathematics there from 2006 to 2010. He also holds secondary faculty appointments in the Tufts University Departments of Physics and Computer Science.

Dr. Boghosian has been a fellow of the American Physical Society since 2000, a recipient of Tufts University’s Distinguished Scholar Award in 2010, a foreign member of the National Academy of Sciences of the Republic of Armenia since 2008, a Fellow of the Jonathan M. Tisch College of Civic Life at Tufts University in 2018, and a Fellow of the Data Intensive Studies Center at Tufts University since 2020.

From 2010 to 2014, while on leave from Tufts University, Dr. Boghosian served as the third president of the American University of Armenia (AUA) in Yerevan, Armenia. During that time, he oversaw the creation, accreditation and inauguration of AUA’s undergraduate program – the first American-accredited bachelor program in the former Soviet Union.

Dr. Boghosian’s interests include mathematical fluid dynamics and kinetic theory, and, more recently, the application of these disciplines to problems in the social sciences, including wealth distribution, wealth inequality, the onset of oligarchy, and opinion dynamics. He is a member of the editorial boards of three scientific journals. He has held visiting positions at, inter alia, the D´epartment de Math´ematiques, Universit´e de Paris-Sud in Orsay, France; the ´Ecole Normale Sup´erieure in Paris, France; Peking University in Beijing, China; University College London; the International Centre for Theoretical Physics in Trieste, Italy; and, most recently, the Economic Research Group at the Central Bank of Armenia.

Dr. Boghosian also maintains a strong interest in international education, and has served on various advisory committees for Tufts University’s abroad programs.

Abstract: Ensuring that every vote is counted as cast is central to the integrity of our democracy, especially for communities that have had to overcome tremendous obstacles to access the ballot box. Technology touches many parts of our elections, and while it can be a useful tool, 

it can also introduce security and privacy concerns and create access issues. Mathematical audits of elections improve transparency, allowing us to appropriately use technology with human oversight. While sham audits have targeted individual communities for scrutiny, audits that rely on solid mathematical principles can provide justified confidence in election outcomes. 

This talk will outline the components of trustworthy elections, including the impact voting technology has on different communities of voters, the risks and concerns that come with different types of voting equipment, and the role of statistical audits of elections in election transparency and trustworthiness. 

Bio: Audrey Malagon is Professor of Mathematics at Virginia Wesleyan University and Mathematical Advisor for Verified Voting, a non-profit, non-partisan organization focused on the intersection of elections and technology. She regularly speaks to audiences about election security and the role of risk-limiting post-election audits and has been quoted in Bloomberg Businessweek, Stateline, and Slate Future Tense. Her media publications include “Our soldiers deserve secure votes” in the Charleston Gazette Mail and “Vote auditing can ensure integrity of elections” in the Virginian Pilot. 

Abstract: Prediction algorithms score individuals, or individual instances, assigning to each one a number in [0,1] that is often interpreted as a probability: What are the chances that this loan will be repaid?  How likely is the tumor to metastasize? What is the probability this person will commit a violent crime in the next two years?  But what is the probability of a non-repeatable event?  Without a satisfactory answer, we cannot even specify the goal of the ideal algorithm, let alone try to build it.

This talk will introduce Outcome Indistinguishability, a desideratum with roots in complexity theory, and situate it in the context of research on the theory of algorithmic fairness.

Bio: Cynthia Dwork, Gordon McKay Professor of Computer Science at the Harvard Paulson School of Engineering, Radcliffe Alumnae Professor at the Radcliffe Institute for Advanced Study, and Affiliated Faculty at Harvard Law School and Department of Statistics, uses theoretical computer science to place societal problems on a firm mathematical foundation.

She was awarded the Edsger W. Dijkstra Prize in 2007 in recognition of some of her earliest work establishing the pillars on which every fault tolerant system has been built for a generation (Dwork, Lynch, and Stockmeyer, 1984). 

Her contributions to cryptography include the launching of non-malleable cryptography, the subfield of modern cryptography that studies -- and remedies -- the failures of cryptographic protocols to compose securely (Dolev, Dwork, and Naor, 1991).  She is a co-inventor of the first secure cryptosystem based on lattices, the current best bet for cryptographic constructions that will remain secure even against quantum computers (Ajtai and Dwork, 1997). More recently, Dwork spearheaded a successful effort to place privacy-preserving analysis of data on a firm mathematical foundation.  A cornerstone of this effort is the invention of Differential Privacy (Dwork, McSherry, Nissim, and Smith, 2006, Dwork 2006), now the subject of intense activity across many disciplines and recipient of the Theory of Cryptography Conference 2016 Test-of-Time award and the 2016 Gödel Prize. Now widely used in industry – for example by Google, MIcrosoft, Uber, and, most prominently, by Apple – differential privacy will also be the foundation of the Disclosure Avoidance System in the 2020 US Decennial Census.

Differentially private analyses enjoy a strong form of stability.  One consequence is statistical validity under adaptive (aka exploratory) data analysis, which is of great value even when privacy is not itself a concern (Dwork, Feldman, Hardt, Pitassi, Reingold, and Roth STOC 2015, and Science Magazine, 2015).

Data, algorithms, and systems have biases embedded within them reflecting designers' explicit and implicit choices, historical biases, and societal priorities. They form, literally and inexorably, a codification of values.  Unfairness of algorithms -- for tasks ranging from advertising to recidivism prediction -- has recently attracted considerable attention in the popular press.  Anticipating these concerns, Dwork initiated a formal study of fairness in classification (Dwork, Hardt, Pitassi, Reingold, and Zemel, 2012).  This is now a thriving subfield of theoretical computer science.

Dwork is currently working in all of these last three areas (differential privacy, statistical validity in adaptive data analysis, and the theory of algorithmic fairness).  Her current principle focus is a complexity-theoretic investigation of the meaning of "individual probabilities."  See the Opportunities tab for more information.

Dwork was educated at Princeton and Cornell.  She received her BSE (with honors) in electrical engineering and computer science at Princeton University, where she also received the Charles Ira Young Award for Excellence in Independent Research, the first woman ever to do so.  She received her M.Sc. and Ph.D. degrees in computer science at Cornell University.

Dwork is a member of the US National Academy of Sciences and the US National Academy of Engineering, and is a fellow of the ACM, the American Academy of Arts and Sciences, and the American Philosophical Society.

Abstract: As algorithms trained via machine learning are increasingly used as a component of screening decisions in areas such as hiring, lending, and education, discussion in the public sphere has turned to the question of what it means for algorithmic classification to be fair to different groups. We consider several of the key fairness conditions that lie at the heart of these debates, and discuss recent research on trade-offs and interventions through the lens of these conditions. We also explore how the complexity of a classification rule interacts with its fairness properties, showing how natural ways of approximating a classifier via a simpler rule can lead to unintended biases in the outcome. 

The talk will be based on joint work with Jens Ludwig, Sendhil Mullainathan, Manish Raghavan, and Cass Sunstein.

Bio:  Jon Kleinberg is the Tisch University Professor in the Departments of Computer Science and Information Science at Cornell University. His research focuses on the interaction of algorithms and networks, the roles they play in large-scale social and information systems, and their broader societal implications. He is a member of the National Academy of Sciences and the National Academy of Engineering, and the recipient of MacArthur, Packard, Simons, Sloan, and Vannevar Bush research fellowships, as well awards including the Harvey Prize, the Nevanlinna Prize, and the ACM Prize in Computing.

Abstract: Recent developments in methods for explaining the decisions of machine learning models have been widely embraced for their ability to provide transparency and accountability without limiting model complexity or compelling model disclosure. Yet applying these methods is far from straightforward and they rarely prove a cure all. This talk identifies a number of unavoidable tensions that decision makers must navigate as they seek to employ these methods—and the deeply subjective judgment that must go into these considerations.

Bio:  Solon Barocas is a researcher in the New York City lab of Microsoft Research and faculty in the Department of Information Science at Cornell University. His research explores ethical and policy issues in artificial intelligence, particularly fairness in machine learning, methods for bringing accountability to automated decision-making, and the privacy implications of inference. He co-founded the ACM conference on Fairness, Accountability, and Transparency (FAccT).

Abstract: Throughout the United States, public school systems seek to incorporate student choices when assigning students to public schools. The design of such school choice systems has been widely studied in the theoretical literature, and academics and practitioners have together developed algorithms for centralized school choice that incentivize families to report truthfully while allowing both choice and district priorities to guide the overall assignment.

In this talk, I will give an overview of key issues that arise when operationalizing and implementing school choice algorithms, and provide technical and computational approaches for addressing some of these issues. While many districts use attendance areas and neighborhood zones to assign students to schools, such tools are typically not considered together with choice, and in joint work with the San Francisco Unified School District we heuristically explore how to best use zones and other school choice policy levers to minimize travel distances while ensuring diversity. Another important operational issue is how to reassign seats between multiple rounds of assignment. We propose and axiomatically justify a class of reassignment mechanisms, characterize the mechanism that maximizes choice while minimizing disruptive reassignment, and show that our optimal mechanism halves reassignment in simulations on NYC public school data. I will end with a reflection on where there continue to be gaps between academic school choice mechanisms and practical implementation, and how we as academics can co-design technical solutions with our non-academic partners.

Bio: Irene Lo is an assistant professor in the department Management Science & Engineering at Stanford University. Her research builds on tools from algorithms and economics to design matching markets and assignment processes, with a focus on public sector and non-profit applications. She is especially interested in resource allocation in education, the environment, and the developing world, and more broadly in socially responsible operations research, and she is a co-organizer of the Mechanism Design for Social Good research initiative. She obtained her Ph.D. from the IEOR department at Columbia University and her A.B. in mathematics from Princeton University.

Abstract: Computer vision has ceased to be a purely academic endeavor. From law enforcement, to border control, to employment, healthcare diagnostics, and assigning trust scores, computer vision systems are being rapidly integrated into all aspects of society. In research, there are works that purport to determine a person’s sexuality from their social network profile images, others that claim to classify “violent individuals” from drone footage. These works were published in high impact journals, and some were presented at workshops in top tier computer vision conferences such as CVPR.

A critical public discourse surrounding the use of computer-vision based technologies has also been mounting. For example, the use of facial recognition technologies by policing agencies has been heavily critiqued and, in response, companies such as Microsoft, Amazon, and IBM have pulled or paused their facial recognition software services. Gender Shades showed that commercial gender classification systems have high disparities in error rates by skin-type and gender, and other works discuss the harms caused by the mere existence of automatic gender recognition systems. Recent papers have also exposed shockingly racist and sexist labels in popular computer vision datasets--resulting in the removal of some. In this talk, I will highlight some of these issues and proposed solutions to mitigate bias, as well as how some of the proposed fixes could exacerbate the problem rather than mitigate it. 

Bio: Timnit Gebru is a senior research scientist at Google co-leading the Ethical Artificial Intelligence research team. Her work focuses on mitigating the potential negative impacts of machine learning based systems. Timnit is also the co-founder of Black in AI, a non profit supporting Black researchers and practitioners in artificial intelligence.  Prior to this, she did a postdoc at Microsoft Research, New York City in the FATE (Fairness Transparency Accountability and Ethics in AI) group, where she studied algorithmic bias and the ethical implications underlying any data mining project. She received her Ph.D. from the Stanford Artificial Intelligence Laboratory, studying computer vision under Fei-Fei Li. Prior to joining Fei-Fei's lab, she worked at Apple designing circuits and signal processing algorithms for various Apple products including the first iPad.

Abstract: Operations research methods have been used to identify and evaluate solutions to the reconfiguration of public school attendance area boundaries for over fifty years. In broad terms, the school redistricting problem seeks to find capacity-feasible assignments of students in a school district to local schools.  This talk will present analysis of the use of operations research for school districting.   The talk will feature a review of the literature, exploring connections between evolving issues in public education and advances in optimization, computing and geographic information systems.  Much of the early work was motivated by Supreme Court decisions to desegregate schools (Brown v. Board of Education, Brown II, Green v. New Kent, Swann v. Charlotte-Mecklenburg).  Around that time, papers appeared in the operations research literature proposing analytical approaches to school desegregation that made use of advances in linear programming.  The talk will examine ways in which these papers modeled the trade-offs between achieving racial balance and minimizing travel distance for students, and the extent to which the resulting analysis impacted policy and court cases.  We will also discuss how the limitations of early models and solution approaches hindered their applicability.  The years since have seen new directions in research to address additional challenges related to the design of school attendance boundaries and leverage emerging advances in optimization, computing, and geographic information systems technology.  The talk will end with a reflection on current issues facing public school districts, including school busing and return-to-school plans amid the COVID-19 pandemic, and the ways in which operations research can be part of these discussions.

Bio: Dr. Karen Smilowitz is the James N. and Margie M. Krebs Professor in Industrial Engineering and Management Science at Northwestern University, with a joint appointment in the Operations group at the Kellogg School of Business.  Dr. Smilowitz is an expert in modeling and solution approaches for logistics and transportation systems in both commercial and non-profit applications, working with transportation providers, logistics specialists and a range of non-profit organizations.  Dr.  Smilowitz is the founder of the Northwestern Initiative on Humanitarian and Non-Profit Logistics.  She has been instrumental in promoting the use of operations research within the humanitarian and nonprofit sectors through the Woodrow Wilson International Center for Scholars, the American Association for the Advancement of Science, and the National Academy of Engineering, as well as various media outlets.  Dr. Smilowitz is an Associate Editor for Transportation Science and Operations Research.   Dr. Smilowitz received the Award for the Advancement of Women in OR/MS from INFORMS and led the winning team in the INFORMS Innovative Applications of Analytics Award.

Abstract: The use of a collection of "neutral", computer generated, redistricting maps is quickly becoming the standard for identifying Gerrymandering and understanding the ways it is achieved.  A number of teams have used these ideas in advising governors and legislatures as well as in a number of court proceedings. The Duke group has been involved in a number of cases which have led to the redrawing of all of the federal and state level redistricting maps for the 2020 elections.

While this will build on the ICERM lecture I will give on the day before, this talk will emphasize more the mathematical structures of the problem and the algorithms we have developed to sample and analyze the space of redistricting. 

More specifically, I will talk both about local Ising-like moves as well as more global moves biased on balanced partitions generated using spanning trees/forests. I will also talk about some recent non-reversable Markov Chain Methods which might be interesting in their own right.  I will also touch on some methods which don't require equilibrium. I will also discuss attempts to visualize the results and understand the geo-political structure of the problem and the different ways gerrymandering manifests.

Overall, I hope to show  you that redistricting is an interesting problem which has generated many interesting mathematical questions already with likely many more to come.

Bio: Jonathan Christopher Mattingly grew up in Charlotte. He graduated from the NC School of Science and Mathematics and received a BS in Applied Mathematics with a concentration in physics from Yale University. After two years abroad with a year spent at ENS Lyon studying nonlinear and statistical physics on a Rotary Fellowship, he returned to the US to attend Princeton University where he obtained a PhD in Applied and Computational Mathematics in 1998 under the supervision of Yakov Sinai. After 4 years as a Szego assistant professor at Stanford University and a year as a member of the IAS in Princeton, he moved to Duke in 2003. He is currently a James B. Duke Professor of Mathematics and a Professor of Statistical Science.

Mattingly's work centers on the long time behavior of random dynamical systems and stochastic partial differential equations in particular. In particular he has definitive works on the ergodic theory of the two-dimensional Navier-Stokes equations. He has also worked on the scaling limits and consistency of various stochastic numerical methods including Markov Chain Monte Carlo and methods to simulate stochastic differential equations. In addition he has worked on a number of biologically motivated problems including fluctuations in cell biochemical networks, the evolution and spread of influenza and the averaging of evolutionary trees.

Since 2013 he has also been working to understand and quantify gerrymandering and its interaction with a region's geopolitical landscape. This has led him to testify in a number of court cases including Common Cause V. Rucho which went all the way to the US Supreme Court. He was also involved with a sequence of North Carolina state court cases  which led to the NC congressional and both NC legislative maps being deemed unconstitutional and replaced for the 2020 elections. He was awarded the Defender of Freedom award by the Common Cause for his work on Quantifying Gerrymandering.

Abstract: Poverty is a multifaceted and dynamic phenomena impacting billions of people worldwide. Despite its prevalence, there remains much to be understood about what makes families susceptible to experiencing economic distress. In recent years, income shocks -- which constitute unexpected expenses or interruptions to one’s income flow -- have garnered increased public attention as being intricately intertwined with poverty. Despite a vast body of empirical work showing the impact of shocks on welfare, they do not play a correspondingly central role in the design of assistance programs. 

In this talk, we present a mathematical and computational analysis of shocks. We pose a set of algorithmic questions about allocation of subsidies in the presence of shocks and present optimal and near-optimal solutions for various general settings. We computationally analyze the impact of shocks on poverty using a longitudinal, survey-based dataset, revealing insights about the interactions of different types of shocks. We discuss how these insights can inform the design and deployment of assistance programs and highlight new directions at this emerging interface between algorithms, public finance, and social work.

Bio: Rediet Abebe is a Junior Fellow at the Harvard Society of Fellows and an incoming Assistant Professor of Computer Science at the University of California, Berkeley. Abebe holds a Ph.D. in computer science from Cornell University and graduate degrees in mathematics from Harvard University and the University of Cambridge. Her research is in artificial intelligence and algorithms, with a focus on equity and justice concerns. Abebe is a co-founder and co-organizer of the multi-institutional, interdisciplinary research initiative Mechanism Design for Social Good (MD4SG). Her dissertation received the 2020 ACM SIGKDD Dissertation Award for offering the foundations of this emerging research area. Abebe's work has informed policy and practice at the National Institute of Health (NIH) and the Ethiopian Ministry of Education. She has been honored in the MIT Technology Reviews' 35 Innovators Under 35 and the Bloomberg 50 list as a one to watch. Abebe also co-founded Black in AI, a non-profit organization tackling representation issues in AI. Her research is influenced by her upbringing in her hometown of Addis Ababa, Ethiopia.

Abstract: There's widespread concern that high-stakes decisions -- made both by humans and by algorithms -- are biased against groups defined by race, gender, and other protected traits. In this talk I'll describe several interrelated threads of research that seek to define, detect, and combat bias in human and machine decisions, drawing on new and old ideas from computer science, statistics, law, and economics. First, in the context of human decisions, I'll demonstrate that one of the most popular statistical tests for assessing discrimination can, in practice, yield misleading results. To address this issue, I'll introduce a new technique, which we call the threshold test, that is designed to circumvent the pernicious problem of infra-marginality. I'll illustrate this method on a novel dataset of nearly 100 million records of traffic stops that we collected from law enforcement agencies across the country. Next, in the context of machine decisions, I'll similarly show that one of the most popular measures of algorithmic bias suffers from serious statistical shortcomings. I'll argue that algorithms built to satisfy this measure can, perversely, harm the very groups they were designed to protect. To demonstrate these ideas, I'll discuss a class of risk-assessment algorithms used by judges nationwide when determining the conditions of pretrial release. I'll conclude by describing one of our recently deployed algorithmic interventions aimed at improving prosecutorial decisions.

Bio: Sharad Goel is an assistant professor at Stanford University in the Department of Management Science & Engineering, with courtesy appointments in Computer Science, Sociology, and the Law School. He's the founder and director of the Stanford Computational Policy Lab, a group that develops technology to tackle pressing issues in criminal justice, education, voting rights, and beyond. In his research, Sharad looks at public policy through the lens of computer science, bringing a new, computational perspective to a diverse range of contemporary social issues, including policing practices, electoral integrity, online privacy, and media bias. Before joining the Stanford faculty, Sharad completed a Ph.D. in applied mathematics at Cornell University, and worked as a senior researcher at Microsoft.