I-DISC Faculty Forum

You are here

Friday, April 14, 2023 | Building C 
9:15AM-12:00 noon (BC210): Faculty Forum
12:00-1:30 PM (BC C3 Lounge): Poster Session & Lunch

Registration is now closed. Please contact Sarah Wing (srw208@lehigh.edu) if you are interested in attending this event. 

This semester, we will be combining the I-DISC Faculty Forum with the External Advisory Council (EAC) meeting. EAC members will give presentations about their research and their organizations.
I-DISC faculty and their students will have an opportunity to showcase their research at a poster session and luncheon. Some members of the EAC will join us in person, and some by Zoom.


9:15 AM
I-DISC Welcome, Overview and Updates 

Presented by I-DISC's Leadership Team: Larry Snyder (ISE), Kate Arrington (Psychology) and Mooi Choo Chuah (Computer Science & Engineering)

9:30 AM
Research Presentation 

"Decoding Nature’s Cookbooks using Many-genome Data" presented by Prof. Wynn Meyer, Biological Sciences
Abtract: What essential items are staples in your pantry? Our genomes are like cookbooks that nature continues to update, leaving in certain essential ingredients and changing others due to chance copying errors or new conditions. Recent and ongoing advances in sequencing technologies have made it easier to read and compare these 3-billion-letter cookbooks across many different populations and species of mammals. Yet reading is not the same as understanding, and genomic researchers are still developing ways to make sense of patterns in these large-scale genomic data. In my work, I compare the contents of many genomes to identify which genetic “ingredients” matter most for mammals’ survival in different environments. These approaches help us to interpret nature’s various recipes and identify components relevant to human and animal health.

10:00 AM Break

10:10 AM
Research Presentation 
"The Trusted AI Strategic Initiative at Sandia National Laboratories" presented by David Stracuzzi, Center for Computing Research, Sandia National Lab
Abstract: Sandia National Laboratories is in the middle of a five-year focused research effort on Trusted AI. The goal is to develop the fundamental insights required to use AI methods in high-consequence national security applications while also improving the practical deployment of AI. This talk summarizes mission drivers for the initiative and presents research thrusts and goals, including a forward-looking discussion of remaining gaps and needs. The talk concludes with a summary of several projects funded by the initiative.

10:40 AM
Research Presentation 
"Leveraging Data Science and Physical Models to Foster Sustainability" presented by Prof. Maryam Rahnemoonfar, Computer Science & Engineering and Civil & Environmental Engineering
Abstract: Sustainable development has become a pressing issue in the modern world, as humanity faces a multitude of environmental challenges that threaten the planet's ecological balance. In this context, data science has emerged as a powerful tool for analyzing and interpreting large volumes of data to gain insights into complex systems, making it an essential component in addressing sustainability challenges.
However, data-driven models are highly dependent on both the quantity and quality of the available labeled data. They also do not generalize well beyond the type of dataset on which they are trained. These AI models are often not explainable, and hence not meaningful for scientific usage as most of the time, they lack consistency with the known laws of physics (e.g. conservation of mass or energy). On the other hand, physical models are based on scientific principles and can easily explain the relationship between input and output variables. However, they are limited in their ability to extract information from the data directly. Moreover, the physical relationships might be too simple or may not include relevant parameters. Models also have relatively coarse resolution and are confined to a specific time interval and region. Using the wealth of available data and knowledge-based models, this talk will explore the ways in which hybrid models can be leveraged to support sustainable disaster management, polar ice monitoring, and language preservation.

11:10 AM Break

11:20 AM
Introduction of I-DISC External Advisory Council Members
Short overview presentations by some of our External Advisory Council members:

  • Mike Liebman, Bloomberg LP: "Current Work Environment & Challenges" 
  • Banu Gemici-Ozkan, Intelligent Systems & Optimization Lab, GE Research
  • Jon Bentley (Bell Labs-retired): "An Annual Master Class in Algorithms"
  • Mike Watson 

12:00-1:30 PM (C3 Lounge area) Lunch provided

Research Group: Robotics / AIRLab

1.  Versatile Modular Multirotor Vehicles.
Jiawei Xu (Graduate student, Computer Science); Diego S. D'Antonio; Brian Zhu; Edward Jeffs; David Saldana, Department of Computer Science & Engineering
Abstract: Aerial vehicles provide unparalleled flexibilities in applications. However, most vehicles are designed to deal with specific tasks and limited ability to physically interaction with the environment. We present a modular vehicle that can adapt to different task requirements. Via reconfiguration of the modules, the vehicles are able to change their actuation capabilities, and find their application in different scenarios.

2. Fully Autonomous Object Transportation using Multiple Robots with Cables
Diego Salazar-D'Antonio (Teaching Assistant, Computer Science & Engineering); Subhrajit Bhattacharya, Department of Mechanical Engineering & Mechanics; David Saldaña Department of Computer Science & Engineering
Abstract: Cables are versatile and lightweight and have the potential to replace heavy grasping mechanisms that are being used in the robotics literature. Although aerial manipulation with cables is lightweight and versatile, current methods in the robotics literature still require human intervention to tie the cable to the object.

3. Flying Swarms
Karen Li ('24 Computer Science); Michael Fitzgerald; Leonardo Santens; Shuhang Hou; Edward Jeffs; Jiawei Xu; Diego S. D’Antonio; David Saldaña, Department of Computer Science & Engineering
Abstract: Our research is about developing low-cost, reliable, lightweight, vision-based autonomous aerial vehicles that interact with physical objects and operate in a realistic environment with the presence of uncertainty and unpredictability.

4. Symbolic Perception Risk in Autonomous Driving
Guangyi Liu (Graduate Student, Mechanical Engineering); Disha Kamale; Cristian-Ioan Vasile; Nadder Motee, Department of Mechanical Engineering & Mechanics
Abstract: We develop a novel framework to assess the risk of misperception in a traffic sign classification task in the presence of exogenous noise. We consider the problem in an autonomous driving setting, where visual input quality gradually improves due to improved resolution, and less noise since the distance to traffic signs decreases. Using the estimated perception statistics obtained using the standard classification algorithms, we aim to quantify the risk of misperception to mitigate the effects of imperfect visual observation. By exploring perception outputs, their expected high-level actions, and potential costs, we show the closed-form representation of the conditional value-at-risk (CVaR) of misperception. Several case studies support the effectiveness of our proposed methodology.

5. Cascading Waves of Fluctuation in Time-delay Multi-agent Rendezvous
Vivek Pandey (Graduate Student, Mechanical Engineering); Guangyi Liu; Christtoforoos Somarakis; Nader Motee, Department of Mechanical Engineering & Mechanics
Abstract: We develop a framework to assess the risk of cascading failures when a team of agents aims to solve the distributed rendezvous problem in time in the presence of exogenous noise and communication time-delay. The notion of value-at-risk (VaR) measure is used to evaluate the risk of cascading failures (i.e., network-based waves of large fluctuations) when some agents have failed to rendezvous. Furthermore, an explicit formula is obtained to calculate the risk of higher-order cascading failures recursively. Finally, from a risk-aware design perspective, we report an evaluation of the most vulnerable sequence of agents in various communication graphs.

Research Group: FinTech/BlockChain and Scalable Systems & Software (SSS)

6. Scaling Zero-Knowledge Proof Generation for Large Blockchain Applications
Tal Derei (Graduate Student, Computer Science); Victor Carolino ('25 Computer Science & Business); Caleb Geren ('25 Computer Science), Jon Klein ('25 IDEAS); Mike Kaufman ('25 Computer Science); Rishad Islam (Graduate Student, Computer Science)
Abstract: Zero-knowledge proofs (ZKPs) are cryptographic primitives that enable a prover to convince a verifier that a statement about some secret is true without leaking information about the secret. We present the scalability and performance of different zero knowledge proving systems that are used in blockchains for scalability and privacy purposes

7. A Blockchain Benchmarking Framework
Jeff Van Buskirk 
(Graduate Student, Computer Science), Nate Cable ('24 Computer Science)
Abstract: We present our approach to a standardized framework for developing blockchain benchmarks. This includes frameworks for workloads, metrics, reporting, and a standardized driver for executing the workload. We then present our implementation of this framework: Blockbench v3, a blockchain benchmark focused on web3 applications.

Research Group: Catastrophe Modeling

8. Application of Triton 2D Hydrodymaniic Flood Model for Lehigh Valley: Hurricane Ida Case
Husamettin Taysi (Graduate Student, Civil Engineering); Y. C. Ethan Yang, Department of Civil & Environmental Engineering
Abstract: This study aims to create a flood model for the Lehigh Valley for the Hurricane Ida which occured in 2021. 5-day of the Hurricane Ida was simulated using inputs such as DEM, streamflow data and two different Manning's n values. Next, results were compared with ground truth data that provided by the USGS. Results showed that Manning's n value of 0.0404 gives more reliable results.This study aims to create a flood model for the Lehigh Valley for the Hurricane Ida which occured in 2021. 5-day of the Hurricane Ida was simulated using inputs such as DEM, streamflow data and two different Manning's n values. Next, results were compared with ground truth data that provided by the USGS. Results showed that Manning's n value of 0.0404 gives more reliable results.

9. Predicting Wildfire Ignition Induced by Conductor-Vegetation Contact Under Strong Winds
Xinyue Wang 
(Graduate Student, Structural Engineering); Paolo Bocchini, Department of Civil and Environmental Engineering
Abstract: Under dry weather conditions that feature high winds, electric power systems have been shown to be a rising source of catastrophic wildfires. The utility-related wildfires are mostly attributed to conductor-vegetation contact which can then lead to flashover (or sparkover) and subsequent ignition. Decision making, such as proactive power shutoffs and vegetation management, can be informed by wildfire risk analysis, in which the ignition probability analysis is a key component. This study focuses on the ignitions caused by the conductor swinging out to nearby vegetation under high winds. The problem is formulated in the context of proactive de-energization with a focus on the transmission system.

Research Group: Mathematical Optimization & Data Science (MODS)

10. Quantum Computing and Optimization
Mohammadhossein Mohammadisiahroudi
(Graduate Student, Industrial & Systems Engineering)
Abstract: Quantum computing has attracted significant interest in the optimization community because it potentially can solve classes of optimization problems faster than conventional supercomputers. Several researchers proposed quantum computing methods, especially Quantum Interior Point Methods (QIPMs), to solve convex optimization problems, such as Linear Optimization, Semidefinite Optimization, and Second-order Cone Optimization problems. Most of them have applied a Quantum Linear System Algorithm at each iteration to compute a Newton step. However, using quantum linear solvers in QIPMs comes with many challenges, such as having ill-conditioned systems and the considerable error of quantum solvers. This research investigates how one can efficiently use quantum linear solvers in QIPMs. Accordingly, Inexact Infeasible QIPM and Inexact Feasible QIPM are developed to solve linear optimization problems. We also discuss how we can get an exact solution by Iterative Refinement without excessive time of quantum solvers. Finally, computational results with a QISKIT implementation of our QIPM using quantum simulators are analyzed.

11. Worst-Case Complexity of TRACE with Inexact Subproblem Solutions for Nonconvex Smooth Optimization
Qi Wang 
(Graduate Student, Industrial & Systems Engineering); Frank E. Curtis; Qi Wang, Department of Industrial & Systems Engineering 
Abstract: An algorithm for solving nonconvex smooth optimization problems is proposed, analyzed, and tested. The algorithm is an extension of the Trust Region Algorithm with Contractions and Expansions (TRACE) [Math. Prog. 162(1):132, 2017]. In particular, the extension allows the algorithm to use inexact solutions of the arising subproblems, which is an important feature for solving large-scale problems. Inexactness is allowed in a manner such that the optimal iteration complexity of O(ε−3/2) for attaining an ε-approximate first-order stationary point is maintained while the worst-case complexity in terms of Hessian-vector products may be significantly improved as compared to the original TRACE. Numerical experiments show the benefits of allowing inexact subproblem solutions and that the algorithm compares favorably to state-of-the-art techniques.

12. A New Multiobjective Heuristic for Creating Political Redistricting Plans While Minimizing Voter Disruption
Brendan Ruskey (Graduate Student, Industrial & Systems Engineering); Lawrence V. Snyder, Deparment of Industrial and Systems Engineering
Abstract: We introduce a multiobjective genetic algorithm (MOGA) for generating political redistricting plans. Unlike all existing MOGAs for redistricting, and most other heuristic algorithms for redistricting, our MOGA produces plans that maximize similarity to an existing plan. Our focus on similarity with a given plan addresses the real-life phenomenon that new redistricting plans are typically created with the previous plan in mind.

Research Group: Explainable Graph Learning

13. Reaction-Diffusion Graph Convolutional Networks: Traffic-Law-Informed Speed Prediction under Limited and Mismatched Data
Yue Sun (Graduate Student, Electrical Engineering); Chao Chen; Yuesheng Xu; Sihong Xie, Department of Computer Science & Engineering; Rick Blum; Parv Venkatasubramaniam, Department of Elecrical & Computer Engineering
Abstract: Traffic speed prediction is crucial for various applications, but existing methods fail when the training data distribution (e.g., traffic patterns on regular days) is limited and different from test distribution (e.g., traffic patterns after a natural disaster). To address this challenge, we propose the traffic-law-informed network called Reaction-Diffusion Graph Convolutional Network (RDGCN), which incorporates a physical model of traffic speed evolution based on a reliable and interpretable reaction-diffusion equation that allows the RDGCN to adapt to unseen traffic patterns. In our experiments, RDGCN demonstrates robustness in handling mismatches and rapid changes. Additionally, RDGCN effectively maintained accuracy by intrinsically imputing missing values. Further, we theoretically proved that under conditions, the RDGCN achieves lower discrepancy in domain generalization.

14. Learning from Biased Implicit Feedback for Unbiased Ranking: From a Causal Perspective
Dan Luo 
(Graduate Student, Computer Science); Brian D. Davison, Department of Computer Science and Engineering
Abstract: We study to leverage user implicit feedback to optimize ranking models through the lens of causality.

Research Group: Machine Learning in Materials Science | The Fredin Group

15. Data Driven Interface Design
Gil Repa
(Graduate Student, Chemistry), Zach Knepp (Graduate Student, Chemistry), Alexandra Howzen (Graduate Student, Material Science & Engineering); Lisa Fredin, Department of Chemistry
Abstract: Material science understanding of epitaxial interfaces are not well understood due to a lack of structural and electronic models that accurately describe known materials. This project creates a new workflow to generate interfacial structures that can be fed into accurate quantum mechanical models in order to fill this need. In addition, the team is working to generate enough data (experimental and computational) to train a neural network to predict new epitaxial interfaces for a range of applications.

Research Group: Quantum Computing

16. On Semidefinite Representations of Second-Order Conic Optimization Problems
Pouya Sampourmahani (Graduate Student, Industrial & Systems Engineering); Mohammadhossein Mohammadisiahroudi; Tamás Terlaky; Department of Industrial and Systems Engineering
Abstract: Second-order conic optimization (SOCO) can be considered as a special case of semidefinite optimization (SDO). In the literature it has been advised that a SOCO problem can be embedded in an SDO problem using the arrow-head matrix transformation. However, a primal-dual solution pair cannot be mapped simultaneously using the arrow-head transformation as we might lose complementarity and duality in some cases. To address this issue, we investigate the relationship between SOCO problems, and their SDO counterpart. Through derivation of standard semidefinite representations of SOCO problems, we introduce admissible mappings. We show that the proposed mappings preserve both feasibility and optimality. Further, we discuss how the optimal partition of a SOCO problem maps to the optimal partition of its SDO counterpart.

17. Characterizing QUBO Reformulations of the Knapsack Problem and Applications to Quantum Computing
Rodolfo Alexander Quintero Ospina 
(Graduate Student, Industrial & Systems Engineering); Luis F. Zuluaga and Tamás Terlaky, Department of Industrial and Systems Engineering
Abstract: It has been shown that quantum computers can outperform classical computers in solving some instances of NP-hard problems, for instance, the Graph partitioning problem, which motivates the use of quantum-based algorithms to solve applied combinatorial problems and more general integer programs. Many of such algorithms need a Quadratic Unconstrained Binary Optimization (QUBO) formulation of the problem, but, most of the literature regarding QUBO reformulations for constrained optimization problems (COPT) is centered around equality constrained problems, and in general, it is not suitable for problems with inequality constraints. Here, we start by focusing on the “simplest” inequality constrained problem, the knapsack problem (KP), and then consider more general integer programs and the algebraic-combinatorial theory needed to obtain exact QUBO reformulations. In particular, we derive different QUBO formulations for the KP, characterize the range of their associated penalty constants, as well as computationally benchmark them through experiments using the quantum approximate optimization algorithm (QAOA) on a gate-based quantum computer. As a byproduct, we correct some erroneous results regarding QUBO reformulations for the KP reported in the literature.

Research group: Bina Lab ( Computer Vision and Remote Sensing Laboratory) 

18. Prediction of Deep Ice Layer Thickness Using Adaptive Recurrent Graph Neural Networks and Physical Features

Ben Zalatan (Graduate Student, Computer Science); Maryam Rahnemoonfar, Departments of Computer Science & Engineering and Civil & Environmental Engineering
Abstract: As we deal with the effects of climate change and the increase of global atmospheric temperatures, the accurate tracking and prediction of ice layers within polar ice sheets grows in importance. Studying these ice layers reveals climate trends, how snowfall has changed over time, and the trajectory of future climate and precipitation. We use an adaptive, recurrent, geometric machine learning model to predict historic snow accumulation by including physical parameters of the ice layers as node features. We found that this model, alongside the inclusion of snow mass balance, meltwater refreezing, and height change due to melt as node features, gives our model consistently high performance.


Friday, November 11, 2022

Research Presentation:

  • "Computational and Human Judgement Forecasts of the Trajectory of Infectious Agents to Support Public Health Decision Making" 
    Tom McAndrew, Computational Scientist, COH

Updates on:

  • DCEJ efforts
  • I-DISC Friendly Review (New Initiative)
  • I-DISC programs & initiatives (fellows, grant writing support, research etc)

Open discussion Q&A

Introductions to Lehigh University Communications and PR members, followed by I-DISCovery on getting help with your website, social media, elevator pitches and media relations.

March 25, 2022


Lehigh Internal Grant Programs 
Overview of VPR's office, and various internal funding opportunities
Presented by Kate Bullard, Research Program Development Officer

I-DISC Fellows Program
Introducing I-DISC Fellows: Dan Luo is an advanced doctoral student with expertise in data science, computation, and machine learning and Alex Sackett is an advanced doctoral student with experience in social psychology, behavioral and cognitive measurement, and a wide range of issues associated with human subjects data collection.
Learn more about the I-DISC Fellows Program and how it can provide a part-time research assistant to help Lehigh faculty with the data science, programming, computational projects and social science aspects of their research.

Nader Motee
 (MEM), will talk about his research and the new AirLab

Opportunities to build diverse UG research groups (Greer Scholars, and others): Josh Agar and Brian Davison will give overview of how I-DISC faculty members could get more involved.

Peer Review Group Initiative 

Other Updates:

  • support for pre- and post-grant awards, projections, 
  • Upcoming events
  • Stay connected: Slack, I-DISC Calendar, You Tube Channel, Twitter

Friday, October 8, 2021


Overview of I-DISC Research Groups
  • FinTech Group
    Presented by Hank Korth, Computer Science and Engineering 
  • Human-Centered Computing (HCC)
    Presented by Eric P.S Baumer, Computer Science and Engineering 
  • Quantum Computing Group
    Presented by Luis Zuluga, Industrial and Systems Engineering 

Lehigh Internal Grant Programs 
Overview of VPR's office, and various internal funding opportunities
Presented by Kate Bullard, Research Program Development Officer

Promoting Equity and Justice in Data and Computing
Overview of I-DISC’s plan for promoting scholarship in equity and justice in data and computing, as well as diversity, equity, and inclusion (DEI).
Presented by Larry Snyder, Kate Arrington, Brian Davison (I-DISC Leadership Team)