Working Groups are an integral part of SAMSI’s research programs, where participants explore research themes determined in the Opening Workshop.
Program on Numerical Analysis in Data Science (Fall 2020)
- Working Group I: Large-scale Inverse Problems and Uncertainty Quantification
The focus of this working group is on advancing computational tools for large-scale inverse problems and uncertainty quantification. We tackle a range of open problems from the development of new regularization approaches for computing solutions to inverse problems to the advancement of technologies for large-scale UQ. We aim to bring together researchers in numerical analysis, probability and statistics, and domain experts with connections to applications of inverse problems.
- Working Group II: Global Sensitivity Analysis
How complex should a computational model be to be useful? Both the mathematical community and domain scientists have long been worried about the lower bound, i.e., minimum complexity. This working group will consider the problem of finding the lowest level complexity sufficient for specific tasks. The working group will gather scientists who are making fundamental contributions to computational modeling under uncertainty through the development of mathematical theory and algorithms for sensitivity analysis of stochastic models as well as goal-oriented dimension reduction for fast UQ and parameter estimation.
- Working Group III: Randomized Algorithms for Matrices and Data
Randomized Numerical Linear Algebra (RandNLA) is an interdisciplinary area which uses randomization as a computational tool for tackling large linear algebra problems. While it has origins in theoretical computer science, RandNLA has connections to many areas of applied and computational mathematics. This working group will bring together researchers in numerical analysis, theoretical computer science, scientific computing, machine learning, statistics, and domain experts with connections to RandNLA. The goal is to not only address foundational questions in RandNLA but to explore and investigate the applications of RandNLA algorithms to new areas in scientific computing and data science including inverse problems, uncertainty quantification, model reduction, and tensor decompositions.
- Working Group IV: Computational Algorithms for Reinforcement Learning
Due to many successful applications in robotics, games, precision health, e-commerce and ride-sharing industries, Reinforcement Learning (RL) has gained great popularity among various scientific fields. The goal of this working group is to join forces of researchers from different academic fields and industry, to explore cutting edge algorithms, applications and theory, including and not limited to deep RL, model-based RL, multi-agent RL, inverse RL and policy evaluation methods.
Online seminars will be presented detailing the latest advances in reinforcement learning applications and theory. The goal is to bring virtual seminars featuring the latest work in applying reinforcement learning methods in many exciting areas (e.g., health sciences, or two-sided markets). Click here for details: https://www.arlseminar.com/
- Working Group V: Dimension Reduction in Time Series
Studies in sufficient dimension reduction for multivariate time series, focusing up Bayesian methods to estimate the central mean subspace, principal component analysis, factor models, and envelope methods.
Program on Combinatorial Probability (Spring 2021)
- Working Group I: Random Simplicial Complex Models (Leader – Sayan Mukherjee, Duke University)
There is an interest in extending random processes on graphs to simplicial complexes. This working group will explore problems in this area
- Working Group II: Group Theory and Representation Theory in Combinatorial Probability (Leaders – Kevin McGoff, UNCC; Evita Nestoridi, Princeton; Langxuan Su, Duke)
The working group will explore the use of ideas from group theory and representation theory in studying the properties of random combinatorial objects. This can include distributions over random groups and using representation theory to study random combinatorial objects such as rankings or Young tableaus.
- Working Group III: Phase Transitions and Algorithms (Leaders – Will Perkins, University of Illinois, Chicago and Nicolas Fraiman, University of North Carolina, Chapel Hill)
This working group will explore the relationship between phase transitions in statistical physics models (e.g. Ising, Potts, hard-core) and efficient algorithms for approximate counting and sampling. While this is a huge field with many different techniques and connections to other areas of mathematics and computer science, we will focus on three specific questions that have emerged in the last few years: 1) when can we develop algorithms that work in the low-temperature, strong interaction regime of a statistical physics model? 2) what are the connections between the three main approaches to approximate counting: Markov chain Monte Carlo, the method of correlation decay, and the polynomial interpolation method? 3) how can we use the worst-case, computational perspective to better understand phase transitions in physics models? The working group will read lecture notes and learn the area as we proceed, so no background in statistical physics is necessary.
- Working Group IV: Symbolic Data Analysis (Leader – Yaser Samadi, Southern Illinois University)
Symbolic data analysis is a relatively new field of study that has been proposed as a general tool for analyzing large and complex datasets by summarizing them into a much smaller and tractable number of features, such as random intervals/rectangles or histograms, each describing a portion of the larger dataset. Unlike classical data, symbolic data have inherent internal variations and structures which must be taken into account; however, classical approaches are not capable of dealing with these issues. The goal is to not only address fundamental questions under the headings of symbolic data analysis but also to explore and develop modern statistical methodologies to analyze symbolic data.
- Working Group I: Big Data Quality – Leaders: Sunshine Hillygus (Duke Political Science and Public Policy), Alex Volfovsky ( Duke Statistics), SAMSI RA: Karen Medlin (UNC Math)
This group will explore methods and frameworks to address quality issues (e.g., incomplete and faulty data) in data-intensive research.
- Working Group II: Misinformation, Information Campaigns, and Event Data – Leaders: Yijyun Lin (University of Nevada, Reno), Ali Hürriyetoğlu, Koç University, Ahmed M. Elmisery, Malm University
This group aims to develop a computational framework to extract useful information from a heterogeneous collection of data sources to improve understanding of causal relations among events and mechanisms stimulate dissemination of misinformation.
- Working Group III: Diffusion of Information in Online Social Networks – Leader: Diego Fregolente, SAMSI
Online social networks are rapidly complementing and even replacing person-to-person social contacts. With hundreds of millions of participants, they have become a major channel for the diffusion of information, and therefore also a place where people are exposed to dangerous manipulations. This group will study the mechanisms and dynamics by which social media are abused and manipulated for the purpose of spreading misinformation.
- Working Group IV: Simulated-data Experimentation to Understand – Leader: Keith O’Rourke, O’Rourke Consulting
This group aims to examine simulation as an alternative mathematical option to better understand the quantitative methods we use. This can include utilizing simulations as a way for one to gain an intuitive understanding of what a method is doing or inform others what a method does or to explore simulation as a thoughtful and rigorous approach to investigate the robustness of a method to analytic assumptions. As a mathematical option it is very general as well as often being self-verifying – simulation can be used to understand simulation. Hands-on working examples based on the methodological interests of the group members will implemented. Challenges of implementation do arise as the number of parameters increase and/or the chosen probability model becomes increasingly complex. The group will start with settings that present few challenges and progress to those with increasing challenges as interests dictate.
- Working Group V: Integrating and Expanding Networks of Networks Theories
Leader – Kate Albrecht, University of Illinois-Chicago
This interdisciplinary group aims to integrate and expand our understanding of dynamics within interdependent networks, network domains, and multiplex networks present in many disparate theories across the sciences. Our group will work to bridge disciplinary perspectives and languages to identify gaps in current theory as well as clarify concepts, operationalizations, and applicable methods to advance new research agendas.
- Working Group VI: Spacial Mediation and Longitudinal Modeling
Leader – Emil Coman, University of Connecticut
Spatial data is increasingly becoming available, for many countries, and the US and individual states, at different spatial precision (counties, towns, Zip, census tract, e.g.). Moreover, such data like Covid positives, mortality, and hopefully soon vaccinations in the US, is being openly posted by health departments. Answering practical questions about why some regions see higher risks, and how such effects happen, through which indirect mechanisms, is still in need for simple analytical and software solutions. When multiple time points of data are added to data, additional questions about how relations change over time, and what causes such changes, are essential. This group examines such simple applied solutions for spatial mediation (and SEM) longitudinal modeling of spatial data.
- Working Group VII: Estimation of Agent-Based Models
Leader – Bruce Rogers
Agent-based models (ABMs) are “bottom-up” descriptions of complex group interactions. We wish to evaluate computational frameworks for estimating the state dynamics of a variety of ABMs.
- Working Group VIII: Cross-Translating Contemporary Causal Modeling Methods
Leaders – Ken Bollen, UNC & Emil Coman, University of Connecticut
When scientists talk about causal inference, each field seems to use their own language. The goal of this working group is to build a ‘Rosetta Stone’ for these discussions so researchers from diverse schools of thought can more easily understand each other.
- Working Group IX: Weights in Data Analysis
Leader – Stas Kolenikov, Abt Associates
Weighted data form the heart of a number of approaches to data analyses in a variety of contexts (weights in complex samples, WLS, propensity score weights, outliers). The group will compare uses of weights and diagnostic tests for their necessity
- Working Group X: Causal Inference with Network Structures
Leader – Weihua An, Emory University
We are interested in studying causal inference problems with underlying network structures. This includes (but is not limited to) problems where we have networks as dependent variables or independent variables, problems that have networked error structures, problems with networked interference/feedback effects, etc.
The focus of this working group is on various kinds of brain network data. This group aims to connect two kinds of researchers: those with biological and psychological questions related to structural and functional brain network data, and those implementing novel mathematical and statistical tools to understand networks
- Working Group XII: Networks and Victimization
Leader – Bernard Coles IV, SAMSI/Duke
Exploring the ways network structure influences individual exposure to violence as well as theories of social learning and crime more broadly.
- Working Group XIII: Networks and Psychology
Leader – Ruchira Datta, Datta Enterprises LLC
Compared to the extensive existing literature using network analysis to understand the brain through a neuroscientific lens, there are noticeably fewer attempts to apply network analysis to psychologic understandings of the mind. We plan to find the intersection of network analysis, psychology, and behavioral game theory to try to model how emotions and internal psychological states move along a social network. REGISTER
Additional information for Working Group Leaders and Web Masters is provided HERE