Abstract: Modern manufacturing generates increasingly rich in-situ sensing data, creating new opportunities for AI to enable automated and intelligent quality control decisions. However, conventional quality control using supervised learning relies on abundant labeled data and assumes that future product defects or process faults resemble those known during training. In practice, new defect types emerge, while abnormal conditions may be rarely observed or completely unknown. These challenges are particularly important for in-situ quality control, where defects must be detected or correctly classified for real-time decision-making, including newly emerging defects with limited or unavailable labels. Moreover, for latent defects that cannot be directly inspected online, defects must instead be predicted from indirect process-sensing signals. This requires mapping process-signal changes to possible defects despite scarce or unavailable defect training samples. This talk explores how advances in AI can address these challenges and enable more adaptive and intelligent in-situ quality control and decision-making for smart manufacturing.
 

Bio: Dr. Judy Jin is the A. Galip Ulsoy Collegiate Professor of Engineering and Professor of Industrial and Operations Engineering at the University of Michigan. Her research lies at the intersection of data science and quality engineering, with a focus on synergistically integrating engineering models, AI, and advanced quality control methods to improve system design and operational performance. She has served as PI/Co-PI on more than $20 million in federally and industry-funded research. Her work has received numerous honors, including 18 Best Paper Awards, the S.M. Wu Research Implementation Award from SME, the Forging Achievement Award from FIERF, the NSF CAREER Award, and the NSF PECASE Award.

Dr. Jin currently serves as Editor-in-Chief of IISE Transactions. She has also served as Vice President of INFORMS, Chair of the INFORMS Quality, Statistics and Reliability Section, and President of the IISE Quality Control and Reliability Engineering Division. She is a Fellow of ASME, IISE, and INFORMS.

Abstract: Many of today's most important applications rely on high-performance manipulation of matrices, tensors, and general arrays of data.  The neural networks at the heart of AI language models are a prime example.  The AI software stack relies on kernels (e.g., for matrix-matrix product) that are specifically designed to deliver throughput close to the hardware's theoretical peak.  However, writing kernels that can deliver maximum performance is extremely challenging.  In this talk, I will discuss some of the work we have done to simplify the task of authoring such kernels.  This includes new abstractions to allow for algebraic manipulation of array layouts and new programming models for more easily expressing the kernels themselves.

Bio: Michael Garland has been a researcher at NVIDIA since 2006.  He joined the company as one of the founding members of NVIDIA Research and is currently Senior Director of Programming Systems Research.  He leads a research group focused on developing technologies that will help programmers take advantage of modern high-performance machines.  Their work spans the software stack with a particular focus on parallel algorithms, programming languages, compilers and runtime systems, and low-level hardware/software interfaces.  He and his team have both published numerous academic articles and contributed to many software packages in broad use by CUDA developers.

Summary: This seminar examines why exponential quantum advantages for NP-hard discrete optimization remain elusive and explores potential explanations and workarounds through the Maximum Cut problem.

Abstract: Quantum computing offers hope for realizing exponential advantages over classical computing. Despite over two decades of work studying quantum approaches for discrete optimization, rigorously provable exponential advantages for approximating optimal solutions to NP-hard problems remain scarce. Why? We will discuss a potential explanation and workarounds, using the Maximum Cut problem as a running example.

Speaker Bio: Ojas Parekh is a theoretical computer scientist who enjoys applying mathematical techniques to practically motivated interdisciplinary problems. He has worked in a variety of fields including discrete optimization, combinatorics, combinatorial scientific computing, and most recently, quantum and neuromorphic computing. A recent passion is helping shape the emerging fields of quantum approximation algorithms and quantum discrete optimization. He co-directs the Quantum Algorithms and Applications Collaboratory (QuAAC) at Sandia National Laboratories and directs the Department of Energy Fundamental Algorithmic Research toward Quantum Utility project (far-qu.sandia.gov), a multi-institutional effort tasked with designing novel quantum algorithms to realize advantages over classical computation, especially for optimization, simulation, and machine learning. Ojas believes the universe loves us because pizza exists.

Abstract: Quantum computing promises to transform how we solve optimization problems—but currently available Noisy Intermediate Scale Quantum (NISQ era) devices alone cannot get us there. A promising path forward lies in hybrid approaches that combine the strengths of classical algorithms with the emerging capabilities of quantum hardware, opening the door to scalable and high-performance solutions. In this talk, I will present recent advances in quantum optimization that bring this vision to life. I will discuss how warm-starting the Quantum Approximate Optimization Algorithm (QAOA) with classical relaxations can dramatically improve performance, how hybrid pipelines leverage combinatorial structure to achieve stronger approximations and noise resilience, and how benchmark families—such as strongly regular graphs and low-degree graphs—help us probe the boundary of quantum advantage. These results connect classical approximation algorithms, convex optimization, and quantum circuit design. Looking ahead, I will outline a research vision for hybrid intelligence, where classical and quantum algorithms operate in synergy rather than competition—rethinking optimization for the next era of computing.

Bio: Swati Gupta is an Associate Professor and the Class of 1947 Career Development Professor at the MIT Sloan School of Management in the Operations Research and Statistics Group. Prior to this, she held a Fouts Family Early Career Professorship as an Assistant Professor at the Stewart School of Industrial & Systems Engineering at Georgia Tech from 2018-2023, where she served as the lead of Ethical AI in the NSF AI Institute on Advances in Optimization from 2021-2023. She received a Ph.D. in Operations Research from MIT in 2017, following a joint Masters and B.Tech in Computer Science from IIT Delhi. Her research bridges optimization, machine learning, and algorithmic fairness, to design algorithms that are both theoretically rigorous and socially impactful, with applications in healthcare, hiring, energy, quantum computing, and beyond. Her work has been recognized by the 2023 NSF CAREER Award, INFORMS Doing Good with OR 2022 (finalist), the JP Morgan Early Career Faculty Recognition in 2021, the NSF CISE Research Initiation Initiative Award in 2019, INFORMS Computing Society 2016 (special recognition), and the INFORMS Service Science Section 2016 (finalist).

Abstract: In this talk, we discuss possible paths toward demonstrating quantum advantage in optimization, focusing primarily on quantum approximate multi-objective optimization. Multi-objective optimization seeks to understand optimal trade-offs between competing objectives by identifying sets of solutions that balance these trade-offs, a task that can be challenging for classical methods. This makes it a natural candidate for exploring potential benefits of quantum computing. We present a quantum algorithm for approximating such trade-offs in combinatorial optimization problems and illustrate its performance using both experiments on IBM quantum hardware and classical simulations. We conclude by discussing complementary paths toward quantum advantage in optimization, highlighting the role of systematic benchmarking efforts such as the Quantum Optimization Benchmarking Library (QOBLIB).
 

Bio: Dr. Stefan Woerner is a Principal Research Scientist, Global Technical Lead for Quantum Optimization, and Manager of the Applied Quantum Science group of IBM Quantum at IBM Research Europe - Zurich. He holds a Master of Science in Applied Mathematics and a Doctor of Sciences in Operations Management from ETH Zurich and is renowned for his work in quantum optimization, quantum finance, and quantum machine learning.

Abstract: This talk surveys the quantum computing stack—from applications and algorithms to hardware controls—highlighting the complexity at each layer and the interdisciplinary skills required to build and operate quantum computers.
 

Bio: Amara Katabarwa earned his Ph.D. from the University of Georgia in 2019 and has since led industry and government-funded efforts in quantum software, benchmarking, error mitigation, and fault-tolerant algorithms, currently serving as a Senior Research Scientist for GTRI.

Abstract: The problem of discriminating among non-orthogonal quantum states arises throughout quantum information processing. For example, in quantum communications, a message is encoded as a quantum state and transmitted over a quantum channel to a receiver that decodes the message. In quantum sensing, a photon in a known state is emitted to gather information about the environment from its reflections. Underlying state discrimination are the design of the quantum measurement device used for measurement and the design of the hypotheses used to encode information or probe the environment. These are the core degrees of freedom in quantum information processing system design. This talk will present the design of three quantum detection tasks, including quantum state and channel discrimination over uncertain channels, sequential unambiguous quantum state and channel discrimination, and quickest quantum change-point detection. These designs make use of a range of classical non-convex optimization methods, including stochastic, bi-level, and multi-objective optimization.
 

Bio: Xiaodong Wang received his Ph.D. in Electrical Engineering from Princeton University. He is a Professor of Electrical Engineering at Columbia University. His research interests span wireless communications, signal processing, machine learning, and quantum information processing. He received the 1999 NSF CAREER Award, the 2001 IEEE Communications Society and Information Theory Society Joint Paper Award, and the 2011 IEEE Communications Society Award for Outstanding Paper on New Communication Topics. He has served as an Associate Editor for the IEEE Transactions on Communications, IEEE Transactions on Wireless Communications, IEEE Transactions on Signal Processing, and IEEE Transactions on Information Theory.

Abstract: It is an exciting time for quantum science and technology research. While the promise of quantum computers continues to motivate billions of dollars of investment by the U.S. government and private sector, quantum sensors have already revolutionized aspects of our daily life. Established technologies like atomic clocks in space underlie GPS and nearly all communication systems that rely on precise time and frequency synchronization. Related technologies like magnetometers are opening up new research frontiers in biomagnetism. At the same time, there remains a cutting edge achieved in research labs but not yet fielded. In this talk, Robert will give a brief overview of quantum sensors and some of their applications. He will discuss the broad topic of magnetic imaging, and the numerous applications that fall within that category. Throughout the talk, he will highlight the interplay between sensor development and signal processing.

Abstract: Two-sided matching platforms, including labor markets, dating apps, accommodation services, and ridesharing systems, must make matching decisions in the presence of choice congestion and strategic platform design challenges. When agents have correlated preferences, popular options can attract too much demand, reduce overall efficiency, and lead to poor market outcomes. In this talk, I will present a framework for two-sided assortment optimization that studies how a platform should decide which options to display to agents and in what order, with the goal of improving matching performance. The main focus will be on maximizing the expected number of matches under general choice models. I will describe several natural classes of platform policies, ranging from static simultaneous displays to fully adaptive sequential policies, and compare their power through adaptivity gap results. I will also discuss polynomial-time approximation algorithms for computing near-optimal policies, and then briefly discuss the revenue-maximization version of the problem, where matches generate pair-dependent rewards. This talk is based on joint works with Alfredo Torrico, Ulysse Hennebelle, and Mohammadreza Ahmadnejadsaein.
 

Bio: Omar El Housni is an Assistant Professor in the School of Operations Research and Information Engineering at Cornell Tech and Cornell University. He is a Field Member of the Center of Applied Mathematics at Cornell. He is also an Amazon Scholar. His research focuses on decision-making under uncertainty where he aims to develop optimization models and design robust and efficient algorithms to address a wide range of operational problems, including revenue management problems such as assortment optimization and online matchings. Omar has spent time as a research scientist at Amazon and Uber where he contributed to the design and implementation of data-driven optimization models for matching and retailing platforms. Omar holds a PhD in Operations Research from Columbia University and an MS and BS in Applied Mathematics from Ecole Polytechnique (Paris). His work has been recognized by INFORMS George Nicholson award and his current research is supported by NSF. 

Abstract: Prior literature shows that self-reports, especially those of sensitive behaviors, tend to be biased. For example, respondents may over-report their family origin and volunteering experiences while under-reporting their drinking and smoking behaviors. In this study I examine a social network-based approach to addressing self-reporting bias, which is to ask peers (e.g. classmates or co-workers) to provide alternative reports that can be used to cross-validate and supplement self-reports. I first formalize the data generation process statistically by proposing a model containing two sub-models, one for the self-reports and the other for the peer-reports. In each of the sub-models, I treat the observed report as a probabilistic realization of a truthful report and a false report. The model can incorporate both self- and peer-characteristics and also dyadic characteristics. A Bayesian approach helps make the model self-containing by imposing a prior on the unknown ego's behavior. Then the Bayesian approach simulates self- and peer-reports according to the data generation process iteratively until the simulated reports emulate the observed reports. Through a case study on self- and peer-reports of smoking among adolescents, I demonstrate that the Bayesian approach can not only reveal how multiple characteristics are related to reporting biases but also incorporate these characteristics effectively into the imputation of the true behavior.
 

Bio: Dr. Weihua An is Professor and Chair of Sociology and Professor of Data and Decision Sciences at Emory University. He is also an affiliated faculty member of the East Asian Studies Program, the Goizueta Business School, and the Rollins School of Public Health. He earned his Ph.D. in Sociology and A.M. in Statistics from Harvard University. Dr. An's research advances theories and methods in network analysis and causal inference. He has published widely in both methodological and substantive journals and also authored multiple statistical packages, which have received 200K downloads in total. He currently serves as the Editor of Sociological Methodology and the Chair-Elect of the Methodology section of the American Sociological Association. He is a recipient of the Faculty Teaching Award from Emory Sociology and the Clifford Clogg Award and the Leo Goodman Award from the American Sociological Association. He has advised 28 PhD dissertations and ten undergraduate honors theses.

Abstract: The agriculture sector is a major economic driver for Africa. However, climatic change is driving severe disruption across Africa, causing prolonged dry seasons and erratic rainfall. As a result, farmers are facing lower farming yield. To mitigate this serious challenge, farmers must adopt modern and green technologies, like solar irrigation. To show the potential of solar irrigation system to mitigate this critical problem, a research project has been developed with the collaboration of Georgia Tech, University of Rwanda and Adama Science and Technology University. The project aims at developing a PV irrigation system that integrates IoT and controllers for real-time monitoring of the performance of the PV-irrigation system and for optimizing water conservation. This project is envisioned to be a prototype model for further scaling up the deployment of PV-irrigation systems in Africa to enhance sustainable agricultural. 
As part of the whole project, this seminar focuses more on the design and economic assessment of the system for a particular village in Rwanda.
 

Bio: Getachew Biru Worku obtained his MSC and PhD in Electrical Engineering from Dresden University of Technology, Germany. He has more than 30 years of academic and research experience in academic institutions and industry. He has extensive experience in teaching a broad range of electrical engineering courses with excellent feedback and acknowledgements from faculties and students. He has been lecturing courses such as Power Quality and Reliability, Power System Operation and Control, High Voltage Engineering, Power System Analysis, Power Transmission and Distribution Engineering, Electrical Machines Design, Smart Grid and Traction Power System in various Ethiopian Universities and University of Rwanda. He is currently employed as an associate professor at the University of Rwanda giving lectures and advising postgraduate students. He is a dedicated researcher in renewable energy deployment focusing on reliability of off-grid and microgrids to bring sustainable electrical power to the underserved areas. His research areas are electrical power and renewable energy and has authored and co-authored more than 45 papers in peer-reviewed journals.

If you cannot make your assigned group day, please feel free to come by on any operating date below: 

• Time: 10:00AM - 3:00PM
• Dates: 
  • 10/6: Staff
  • 10/7: Faculty
  • 10/8: Ph.D. Students

Recommendations for attire: 

• Wear blue, or dark-colored clothing (will be on a white backdrop)
• Do NOT wear red
• Avoid wearing large jewelry and patterns

Abstract: Former Delta data scientist Nick Sovich introduces SCORPION, a machine learning model that predicts standby boarding probabilities for nearly 500,000 Delta flights up to four months in advance. ​

Built from his experience on 500+ standby flights, SCORPION continuously retrains on historical outcomes and is now used by thousands of Delta employees.​

Abstract: This talk describes the progress in analysis and optimization of Markov Decision Processes (MDPs) and Partially Observable MDPs (POMDPs) with infinite state spaces and possibly noncompact action sets. We shall also discuss applications to inventory control and to controlled linear Gaussian systems. 
 

Bio: Eugene A. Feinberg received MS in Applied Mathematics and Computer Engineering from Moscow University of Transportation, Russia, in 1976 and Ph.D. in Probability and Statistics from Vilnius University, Lithuania, in 1979. Currently he is Distinguished Professor at the Department of Applied Mathematics and Statistics of Stony Brook University.

His research interests include stochastic models of operations research, probability theory, real analysis, Markov Decision Processes, and applications of operations research and statistics to engineering, biology, and medicine. He has published more than 100 papers and edited the Handbook on Markov Decision Processes. His research has been partially supported by the National Science Foundation, Office of Naval Research, National Institute of Health, New York Office of Science, Technology and Academic Research, and private industry. He has served as a Council Member of the INFORMS Applied Probability Society and on several editorial boards. He is a fellow of INFORMS.

Abstract: Digital twins promise continuous, high-fidelity representations of physical systems, yet their practical deployment is often constrained by the computational cost of simulation and calibration. This talk presents a unified perspective on making digital twin calibration more efficient, adaptive, and scalable by learning where computational effort matters most.

The first part addresses offline calibration of block-structured models. I introduce a doubly importance-driven calibration method that learns which parameter blocks and which observations are most informative, and directs simulation effort toward the parts of the problem that contribute most to the optimization. This targeted allocation substantially reduces computational cost while preserving calibration accuracy.

The second part turns to online calibration, where model parameters evolve over time and computational resources are limited. The proposed framework combines fast, edge-side surrogate filtering with periodic, cloud-side discrepancy correction based on high-fidelity simulations. This two-tier architecture enables real-time parameter adaptation while maintaining consistency with the underlying digital twin.

Together, these methods show how learning what matters—in parameters, observations, and computation—can support digital twins that remain accurate, scalable, and continuously adaptive in practice.
 

Bio: Dr. Eunshin Byon is a Professor in the Department of Industrial and Operations Engineering at the University of Michigan, Ann Arbor, where she also serves as Director of the Master's Program. She received her Ph.D. in Industrial and Systems Engineering from Texas A&M University in 2010. Her research spans data science, digital twin modeling and analysis, and quality and reliability engineering, with applications in energy, healthcare, and manufacturing systems. She served as Chair of the Quality, Statistics, and Reliability (QSR) Section of INFORMS in 2019–2020, and her research group has received multiple research and teaching awards from INFORMS, IISE, and IEEE. Dr. Byon is currently a Senior Editor for the INFORMS Journal on Data Science (2024–present) and a Department Editor for IISE Transactions (2021–present). She previously served as an Associate Editor for IISE Transactions (2019–2021), the INFORMS Journal on Data Science (2020–2024), and IEEE Transactions on Automation Science and Engineering (2019–2021).

Abstract: Recent advances in neuroimaging have produced massive and heterogeneous datasets, ranging from fMRI with high spatial resolution to EEG with high temporal resolution, characterized by complex spatiotemporal correlations and substantial inter-subject variability. Traditional regression models and Gaussian process (GP) approaches with fixed parametric kernels often fail to model such complex data effectively while maintaining scalability and interpretability. This talk introduces a family of modern Bayesian GP frameworks that integrate deep kernel learning, neural network priors, and geometric modeling for large-scale neuroimaging analysis. An example is the Deep Kernel Learning Process (DKLP), which embeds deep neural networks within GP priors to learn data-adaptive covariance structures directly from imaging data. DKLP provides a unified modeling foundation for image-on-scalar, scalar-on-image, and image-on-image regression, supported by theoretical guarantees and efficient posterior computation. Applications to fMRI data from the Adolescent Brain Cognitive Development (ABCD) study reveal reproducible cortical activation patterns associated with cognitive ability, while analyses of EEG-based brain–computer interface data demonstrate robust neural decoding under high noise. I will also discuss scalable heat-kernel GPs on manifolds and thresholded GP–based spatially varying neural network priors, which together expand the scope of Bayesian inference for complex neuroimaging data.
 

Bio: Dr. Jian Kang is Professor and Associate Chair for Research in the Department of Biostatistics at the University of Michigan. His research lies at the intersection of Bayesian statistics, machine learning, and artificial intelligence, with applications in neuroimaging, brain–computer interfaces, omics, and precision medicine. He has published more than 175 papers in leading statistics, machine learning, and biomedical journals. Dr. Kang has served as an Associate Editor for several premier statistical journals, including the Journal of the American Statistical Association (JASA), The Annals of Applied Statistics (AOAS) and Biometrics. He is a Fellow of both the Institute of Mathematical Statistics (IMS) and the American Statistical Association (ASA). He currently serves as Chair of the ASA Section on Statistics in Imaging.

The H. Milton Stewart School of Industrial and Systems Engineering at Georgia Institute of Technology is pleased to welcome you to Atlanta, Georgia for Georgia Statistics Day 2026.

Georgia Statistics Day is an annual event that promotes interdisciplinary statistics research across the University of Georgia, the Georgia Institute of Technology, and Emory University, with the venue rotating among the participating institutions. The 2026 event will feature a keynote lecture, two semi-plenary lectures, parallel research sessions, a student poster session, and ample opportunity for exchange between academia and industry.

We are honored to have Prof. Jianqing Fan, Frederick L. Moore ’18 Professor of Finance and Professor of Statistics, Machine Learning, and Operations Research and Financial Engineering at Princeton University, as our keynote speaker. The semi-plenary speakers are Prof. Sivaraman Balakrishnan from Carnegie Mellon University and Prof. Mladen Kolar from the University of Southern California.

This one-day workshop brings together faculty, students, and industry researchers from across Georgia and the Southeast for invited talks, interdisciplinary exchange, mentoring, and networking. Registration and poster submissions are open through September 28, 2026.
 

Georgia Statistics Day 2026 website

• Complete the five challenges below.
• Each challenge is worth 10 points.
• Complete all five on September 3rd  for a 10-point bonus.
• 40 points = Ice Cream Ticket
• 50 points = ISyE Swag Giveaway Drawing Entry Ticket
• 60 points = Grand Prize Drawing Entry Ticket

**Please note that Challenges 2, 3, and 4 will be available only on Thursday, September 3, 11am-12:15pm only.

Challenge 1:  Photo Quest (Worth: 10 Points)

Capture George Tower, study spaces/Collaborative Learning Space(s), student organizations, faculty member, at least three new ISyE students, CASE team member(s), Academic Office team member(s), XR Makerspace, and what you would consider the best view from the building.  Upload and tag gt_isye on Instagram to receive points for this challenge. See attached flyer for more information on how to upload. 

Challenge 2:  Club Crawl (Worth: 10 Points)

Visit Alpha Phi Mu, BISyE, IISE, Student Ambassadors, Supply Chain& Logistics Club, and WISyE.

 Challenge 3:  Resource Rally (Worth: 10 Points)

Visit Academics (Advising, OR Master’s in IE, OR MSA, OR, PhD tables), CASE, ISyE XR Makerspace, and Faculty Undergraduate Research labs.

Challenge 4:  Did You Know? Discovery (Worth: 10 Points)

Collect fun facts from offices, labs, and organizations that you visit. You can continue to visit the Academic Office (houses Advising, Master’s and PhD in IE, MSA, and CASE) after the “kickoff” to find your “fun fact” throughout the week.

Challenge 5:  Leave Your Mark (Worth: 10 Points)

Complete the Achievement Board AND answer a prompt on the ISyE Wall.