# Calendar

During the fall 2018 semester, the Computational Social Science (CSS) and the Computational Sciences and Informatics (CSI) Programs have merged their seminar/colloquium series where students, faculty and guest speakers present their latest research. These seminars are free and are open to the public. This series takes place on Fridays from 3-4:30 in Center for Social Complexity Suite which is located on the third floor of Research Hall.

If you would like to join the seminar mailing list please email Karen Underwood.

COLLOQUIUM OF THE COMPUTATIONAL MATERIALS SCIENCE CENTER

AND THE DEPARTMENT OF COMPUTATIONAL AND DATA SCIENCES

(CSI 898-Sec 001)

**Scientific Data Mining and Its Applications in Tropical Cyclone Research**

Ruixin Yang, Associate Professor

Department of Geography and Geoinformatics

George Mason University

Fairfax, VA

October 31, 4:30 pm

Exploratory Hall, Room 3301

Fairfax Campus

In this talk I will give an overview of of scientific data mining (SDM) followed by a few examples of SDM applications to the analysis of tropical cyclones (TCs), focusing on their intensity changes. Because rapidly intensifying (RI) tropical cyclones are the major error sources in TC intensity forecasting, association rules facilitate the RI process by mining for sets of conditions that have strong interactions with rapidly intensifying TCs. The technique of association rules explores associations among multiple conditions in a simple manner identifying a predictor set with fewer factors but improved RI probabilities. Furthermore, in searching the “optimal” RI condition combinations, a peculiar condition combination was identified that gives a very high RI probability. Such combination can be considered as a sufficient condition for RI that almost guarantees that a RI will take place. Applications of classification techniques to the intensity forecasting will also be discussed. Several drawbacks and future directions for SDM with the TC intensity change problem will be discussed at the end of the talk.

Refreshments will be served at 4:15 PM.

———————————————————————-

Find the schedule at http://www.cmasc.gmu.edu/seminars.htm

COLLOQUIUM OF THE COMPUTATIONAL MATERIALS SCIENCE CENTER

AND THE DEPARTMENT OF COMPUTATIONAL AND DATA SCIENCES

(CSI 898-Sec 001)

**Robust Estimation of Value-at-Risk for Quantitative Risk Management: Applications to Climatology, Insurance, Accidents and Financial Analysis**

Sabyasachi Guharay

U.S. Department of the Treasury

Washington, D.C.

November 7, 4:30 pm

Exploratory Hall, Room 3301

Fairfax Campus

Establishing robust quantitative metrics which allow decision makers to determine the amount of risk in a system with extreme loss events is a problem of interest in many scientific fields. One of the fundamental metrics which is universally accepted in all fields of risk management is the quantity known as Value-at-Risk (VaR). A subfield of risk management, modern Operational Risk Management (ORM), closely investigates methodologies on robustly estimating VaR, “Robust Estimation of VaR.” Currently, academic researchers and industry practitioners are actively looking at ways to make this estimate more statistically robust and accurate with minimal assumption requirements.

In this talk I will present two new quantitative approaches for estimating VaR that are agnostic regarding the relationship between frequency and severity: (1) Data Partition of Frequency and Severity (DPFS) using K-means to estimate VaR; (2) Distribution based partitioning (DBP) of frequency and severity using copulas. Verification is conducted on five simulated scenario datasets while validation is conducted on five publicly available datasets from four different domains: –US Financial Indices data of Standard & Poor’s 500 and Dow Jones Industrial Average; –Chemical Loss spills as tracked by the US National Coast Guard; –Australian automobile accidents; –US hurricane data. It is observed that previous VaR calculations inaccurately estimate the VaR for 80% of the cases in simulated data and 60% of the cases in real-world data studies while new methodologies attains accurate VaR estimates which are within the 95% confidence interval bounds for both simulated and real-world data.

Refreshments will be served at 4:15 PM.

———————————————————————-

Find the schedule at http://www.cmasc.gmu.edu/seminars.htm

COLLOQUIUM OF THE COMPUTATIONAL MATERIALS SCIENCE CENTER

AND THE DEPARTMENT OF COMPUTATIONAL AND DATA SCIENCES

(CSI 898-Sec 001)

**Using Coarse-Grained Models to Access Expanded Length and Time Scales for Nanoscience Applications**

K. Michael Salerno, Jr.

Center for Computational Materials Science

Naval Research Laboratory

Washington D.C.

November 14, 4:30 pm

Exploratory Hall, Room 3301

Fairfax Campus

Polymers and other soft materials have an important role as a coating for nanoscale building blocks like metallic nanoparticles (NPs) and nanorods. This coating mediates interactions between these building blocks and their environment. Atomistic molecular dynamics (MD) simulations are ideal for examining the role of chemistry and atomic interactions at the sub-nanometer scale, for example in the interactions between a NP and a solvent, or between pairs of NPs. Unfortunately, atomistic MD simulations are limited to lengths of order 50 nm and times of order 50 ns. The time scale limitation precludes modeling nanoscale self-assembly, and limits dynamic simulations to extremely high rates of deformation or thermalforcing. These simulations are also limited to sizes that represent a small number of NPs, making it impossible to model large assembled structures.

Faced with these limitations, we have developed coarse-grained (CG) models of polyethylene, a simple polymer used to coat NPs and nanorods. These models have enabled simulations of bulk polymer melts that overcome the limits of atomistic MD by providing a computational speedup of greater than 104 while retaining fundamental details at the sub-nanometer scale. These details produce the viscoelastic properties and semi-crystalline behavior that are intrinsic to polyethylene and that are missed by generic CG models. When applied to a NP coating the CG models capture the coating morphology, indicating the value of using these CG models in nanoscale applications.

Refreshments will be served at 4:15 PM.

———————————————————————-

Find the schedule at http://www.cmasc.gmu.edu/seminars.htm

COLLOQUIUM OF THE COMPUTATIONAL MATERIALS SCIENCE CENTER

AND THE DEPARTMENT OF COMPUTATIONAL AND DATA SCIENCES (

CSI 898-Sec 001)

**Topological Quantification of Microstructures**

Tom Wanner

Department of Mathematical Sciences

George Mason University

Fairfax, VA

November 21, 2016, 4:30 p.m.

Exploratory Hall, Room 3301

Fairfax Campus

Many applied processes generate complex microstructures or patterns which are hard to quantify due to the lack of any underlying regular structure. These patterns may evolve with time or include some element of stochasticity. The resulting variations in the detail structure frequently force one to concentrate on rougher geometric features. From a mathematical point of view, several notions from algebraic topology suggest themselves as natural quantification tools in such a setting. In this talk I will describe some of these tools, in particular homology and persistent homology, and how they can be efficiently computed using open source software. I will also present some applications motivated by materials science problems.

Refreshments will be served at 4:15 PM.

———————————————————————-

Find the schedule at http://www.cmasc.gmu.edu/seminars.htm

**COMPUTATIONAL RESEARCH AND APPLICATIONS SEMINAR **

Oscar Olmedo, PhD

Data Scientist

CACI International, Inc.

**Calibration for Probabilistic Classification**

Monday, August 28, 4:30-5:45

Exploratory Hall, Room 3301

ABSTRACT: This talk will be a review of calibration methods for classifiers that make probabilistic predictions on a scale 0 to 1. It is known that certain classification methods, such as Naïve Bayes or Random Forest make biased predictions that to not match the true posterior probabilities. By calibrating the predictions made by classifiers the true probability of the predicted class can be determined. This type of calibration can be crucial for real-world decision making problems in medicine, business, marketing, and finance. In this talk I will focus on applications in marketing.

Part two: Marketing yourself for future careers outside of academia

It is known that the number of jobs in academia is not rising as fast as the number of PhD’s graduating. Currently a new career option is available to these PhDs, the Data Scientist. But how does one make the transition out of academia to this hot new field? I will discuss strategies for marketing yourself as well as tools necessary to be successful in your transition.

Dr. Oscar Olmedo is an alumnus of George Mason University who studied physics in undergrad (2004), Computational Sciences and Informatics Masters (2007), and Computational Sciences and Informatics PhD (2011) with a concentration in solar physics under Dr. Jie Zhang. After graduating in 2011, Dr. Olmedo went on to NRL as an NRC fellow for two years, and briefly worked at NASA Goddard for a few months in 2013 before moving to Syntasa, a startup focusing on ecommerce/marketing analytics. In 2015, he moved to CACI to work on cyber security research as a DARPA contractor.

A copy of Dr. Olmedo’s presentation is found here: OLMEDO_PRESENTATION_8.28.17

**COMPUTATIONAL RESEARCH AND APPLICATIONS SEMINAR**

Claudio Cioffi-Revilla, Professor

Computational Social Science

Department of Computational and Data Sciences

Director, Center for Social Complexity

George Mason University

**Computational Modeling of Terrorism**

Monday, September 11, 4:30-5:45

Exploratory Hall, Room 3301

Computational social scientists have investigated terrorism for decades, but only recently has the field advanced to creating the first testable formal theories. This talk will review some important background and present recent advances in agent-based modeling of terrorism, based on radicalization theory and research. Enduring challenges will also be covered, as opportunities for research projects, theses, and dissertations.

Dr. Cioffi-Revilla is a Professor of Computational Social Science, founding and former Chair of the Department of Computational Social Science, and founding and current Director of the Mason Center for Social Complexity at George Mason University. He holds two doctoral degrees in Political Science and International Relations and his areas of special interest include quantitative, mathematical, and simulation models applied to complex human and social systems.

**COMPUTATIONAL RESEARCH AND APPLICATIONS SEMINAR **

Jim Simpson, PhD

Cybersecurity and Data Science Consultant

**Ages and Lifetimes of U.S. Firms: Why Businesses Should NOT be Treated Like People**

Monday, September 18, 4:30-5:45

Exploratory Hall, Room 3301

Abstract: An overview of Deep Learning with quick introductions to techniques such as Word Embeddings, Autoencoders, Transfer Learning, Attention Models, and Generative Adversarial Networks.

Dr. Simpson is a seasoned researcher in the areas of machine learning, deep learning, and cybersecurity. He has served as Principal Investigator and Data Scientist on several DARPA programs with a focus on machine learning applications to data fusion, prediction, and unsupervised anomaly detection problems. He holds a Ph.D. in Electrical Engineering from North Carolina State University where he developed novel receivers and algorithms for undersea communications.

**COMPUTATIONAL RESEARCH AND APPLICATIONS SEMINAR **

Lauren Deason, PhD

Lead Data Scientist

Punch Cyber Analytics Group

**Detecting**

**Automated and Coordinated Activity in Cyber Logs at Scale**

Monday, September 25, 4:30-5:45

Exploratory Hall, Room 3301

ABSTRACT: This presentation will detail novel analytic methods for processing time series data at scale using techniques drawn from digital signal processing and document matching. These methods can be applied to detect coordinated and automated cyber activity, to match patterns present in time series data, and to fuse together disparate datasets.

Dr. Deason holds a BS in Applied Mathematics from the University of Virginia, a MA in Mathematics with an emphasis in Real Analysis and Probability Theory from UC Berkeley, and a PhD in Economics from the University of Maryland, College Park. Dr. Deason has 10 years of experience in mathematical modeling and data science, spanning employment as a Professor of Mathematics, an Economist, and a Data Scientist. Dr. Deason’s past experience includes developing dynamic stochastic models within a game theoretic framework to explore the effects of trade policy uncertainty as well as estimating empirical models to explain various phenomena. More recently, Dr. Deason has developed multiple algorithms for detecting and classifying periodic and coordinated behavior in a variety of contexts on large data sets as part of DARPA’s Network Defense Program.

**Computational Social Science**

Robert Axtell, PhD

Computational Social Science Program, Department of Computational and Data Sciences,

College of Science

Department of Economics, College of Humanities and Social Sciences

Krasnow Institute for Advanced Study

George Mason University

Co-Director

Computational Public Policy Lab

Krasnow Institute for Advanced Study and Schar School of Policy and Government

External Professor, Santa Fe Institute (santafe.edu)

External Faculty, Northwestern Institute on Complex Systems (nico.northwestern.edu)

Scientific Advisory Council, Waterloo Institute for Complexity and Innovation (wici.ca)

**Getting Younger by Growing Older: U.S. Firms Gain Longevity as they Age**

Friday, September 29, 3:00 p.m.

Center for Social Complexity Suite

3rd Floor, Research Hall

Abstract: Using data on the entire population of American firms I will first show that the distribution of firm ages is approximately stationary, with small ‘defects’ arising at the start of last decade’s Financial Crisis now propagating through the distribution. From these data I will derive the distribution of U.S. firm lifetimes and demonstrate that it has a specific structure that conforms to economists’ intuitions about new (and small) firms having higher failure probabilities than older (and larger) firms. I will then demonstrate that the large body of statistical theory on ‘survival analysis’ is directly applicable to firms. Specifically, I will focus on firm hazard functions and empirically show that for the U.S. this is a power law over a wide range of ages, ostensibly a new finding, declining with age. This permits computation of the expected remaining lifetime of firms as a function of their age, an INCREASING function, implying that American firms gain longevity as they get older, a very non-biological type of aging. Conditioning on firm size produces further results. Specifically, using the Cox ‘proportional hazards’ specification, the reduction in failure probability associated with larger size is quantified. At the end I will demonstrate that an ABM of firm dynamics can be calibrated to reproduce all of these features of U.S. firms.

Rob Axtell earned an interdisciplinary Ph.D. degree at Carnegie Mellon University, where he studied computing, social science, and public policy. His teaching and research involves computational and mathematical modeling of social and economic processes. Specifically, he works at the intersection of multi-agent systems computer science and the social sciences, building so-called agent-based models for a variety of market and non-market phenomena.

His research has been published in the leading scientific journals, including *Science *and the *Proceedings of the National Academy of Sciences*, *USA*, and reprised in *Nature*, and has appeared in top disciplinary journals (e.g., *American Economic Review*, *Computational and Mathematical Organization Theory*, *Economic Journal*), in general interest journals (e.g., *PLOS One*) and in specialty journals (e.g., *Journal of Regulatory Economics*, *Technology Forecasting and Social Change*.) His research has been supported by American philanthropies (e.g., John D. and Catherine T. MacArthur Foundation, Institute for New Economic Thinking) and government organizations (e.g., National Science Foundation, Department of Defense, Small Business Administration, Office of Naval Research, Environmental Protection Agency). Stories about his research have appeared in major magazines (e.g., *Economist*, *Atlantic Monthly*, *Scientific American*, *New Yorker*, *Discover*, *Wired*, *New Scientist*, *Technology Review*, *Forbes*, *Harvard Business Review*, *Science News*, *Chronicle of Higher Education*, *Byte*, *Le Temps Strategique*) and newspapers (e.g., *Wall St. Journal*, *Washington Post*, *Los Angeles Times*, *Boston Globe*, *Detroit Free Press*, *Financial Times*). He is co-author of *Growing Artificial Societies: Social Science from the Bottom Up* (MIT Press) with J.M. Epstein, widely cited as an example of how to apply modern computing to the analysis of social and economic phenomena.

**COMPUTATIONAL RESEARCH AND APPLICATIONS SEMINAR **

Kirk Borne, PhD

Principal Data Scientist and an Executive Advisor

Booz Allen Hamilton

Monday, October 2, 4:30-5:45

Exploratory Hall, Room 3301

ABSTRACT: Smart data are essential when faced with massive-scale data collections. “Smart” refers to data that are tagged or indexed with meaning-filled metadata that carry information about the semantic meaning of the data, its applications, use cases, content, context, and more. Such meta-tags enable efficient and effective discovery, description, and delivery of the right data at the right time, both to humans and to automatic processes.

Dr. Borne advises and consults with numerous organizations, agencies, and partners in the use of data and analytics for discovery, decision support, and innovation. Previously, he was Professor at George Mason University (GMU) for 12 years in the CSI and CDS programs, where he did research, taught, and advised students in data science. Prior to that, Dr. Borne spent nearly 20 years supporting data systems activities on NASA space science research programs, including a role as NASA’s Data Archive Project Scientist for the Hubble Space Telescope.

Recently, Dr. Borne was ranked #2 worldwide among all Big Data experts to follow. http://ipfconline.fr/blog/2017/05/22/fine-list-of-50-top-world-big-data-experts-to-follow-in-2017-with-moz-social-score/