Skip to main content

TrAC AI Seminar Series

 

Spring 2023--  TrAC semianr talks can be found on TrAC home page at   

https://trac-ai.iastate.edu/events/category/seminars/

TrAC Seminar Series Fall 2022 

Translational AI Center has 40 Iowa State artificial intelligence faculty members and subject matter experts.

The TrAC seminar meets Fridays at the lunch time, with a lead speaker, raising issues and exchanging ideas on topics of current interest in the area of Translational AI and related advances. The format consists of a lead presentation of about 50 minutes, followed by questions and discussions. 

9/2 - All hands meeting

9/9 - Hailiang Liu [Math] on ``Data-driven optimal control with neural network modeling of gradient glows" flyer

9/23 -Henri Chung on "Predicting Antimicrobial Resistance with Machine Learning: A Cautionary Tale of Missing Data” flyer 


TrAC Seminar Series Spring 2022 

Meeting times: Friday(s), 12:00--1:00pm (CST) 
Talk: 12:00 --12:50:00; Q&A: -- 12:50--1:00pm
Zoom link:  
Organizers: Hailiang Liu  hliu@iastate.edu, Soumik Sarkar,  Baskar Ganapathysubramanian 
TrAC contact: Aditya Balu baditya@iastate.edu    

The Translational AI Center for Research and Education (TrAC) is a newly established ISU center, a forum for integration of the activities and skill-sets of multiple ISU investigators with external collaborators to accelerate knowledge transfer from fundamental scientific advances in AI to industrial applications. The TrAC seminar meets Fridays at the lunch time, with a lead speaker, raising issues and exchanging ideas on topics of current interest in the area of Translational AI and related advances.  The format consists of a lead presentation of about 50 minutes, followed by questions and discussions.
 

Jan 21 - All hands meeting

Jan 28 - AIIRA/COALESCE joint seminar

Feb 4  Hailiang Liu [Math] on " Mathematical Problems in Deep Learning" flyer 

Feb 11 Jue Yan [Math] on "Cell-average based neural network fast solvers for time dependent partial differential equations"  flyer 

Feb 18  Shana Moothedath [ECE] on ``A Game and Control Framework for Modeling and Mitigating Advanced Persistent Threats on Cyber-Physical Systems" flyer  

Feb 25  AIIRA /COALESCE joint seminar:  
Mark Ryan [Wageningen Economic Research] on ``Ethical and Societal Considerations for the Development and Use of Agricultural Robots" flyer 

Mar 4 Cody Fleming [Mech] "Towards Autonomous Systems That Leverage First Principles and Black Boxes" flyer 

Mar 11 Alicia Carriquiry [STAT] ``Machine Learning in Forensic Applications" flyer

Mar 25 AIIRA/COALESCE joint seminar

April 1 Shayok Chakraborty [CS, Florida State University] ``Learning with Weak Supervision: Algorithms and Applications" flyer 

April 15 Volkan Isler "From surveying frames to tidying our homes with robots" flyer 

TrAC Journal club:  April 15 at 3pm [CST] https://iastate.zoom.us/j/93596996347 
Xuping TIAN [Math]  "Adaptive gradeint method with energy and momentum" flyer 

April 22-23: Friday (8:45am-5pm) and Saturday (9am-12:30pm)

TrAC workshop "Scientific Marchine Learning: Foundations and Applications" 

This workshop sought to bring together top experts from areas of scientific machine learning to discuss progress that has been made on scientific machine learning research, and to identify promising avenues where theory is possible and useful. There will be several invited talks each day and also spotlight talks by young researchers. This meeting will expose participants to some of the main current trends and recently developed tools in scientific machine learning research and applications.

An updated schedule can be found at

https://trac-ai.iastate.edu/Activities/workshops/SciML2022.html

Day 1 pre-lunch sessions are going to happen at 2206 Student Innovation Center

Day 1 post-lunch sessions with lightning talks and hands-on ScML tutorial are at 4202 Student Innovation Center

Day 2 sessions with four invited talks are at 0114 Student Innovation Center

 

Plenary talk: April 22, 9:00-10:00am and three invited talks are at 2206 Student Innovation Center

George Em Karniadakis
Brown University

From PINNs to DeepOnet: Two Pillars of Scientific Machine Learning
We will review physics-informed neural network and summarize available extensions for applications in computational mechanics and beyond. We will also introduce new NNs that learn functionals and nonlinear operators from functions and corresponding responses for system identification. The universal approximation theorem of operators is suggestive of the potential of NNs in learning from scattered data any continuous operator or complex system. We first generalize the theorem to deep neural networks, and subsequently we apply it to design a new composite NN with small generalization error, the deep operator network (DeepONet), consisting of a NN for encoding the discrete input function space (branch net) and another NN for encoding the domain of the output functions (trunk net). We demonstrate that DeepONet can learn various explicit operators, e.g., integrals, Laplace transforms and fractional Laplacians, as well as implicit operators that represent deterministic and stochastic differential equations. More generally, DeepOnet can learn multiscale operators spanning across many scales and trained by diverse sources of data simultaneously.

 

Levon Nurbekyan 
University of California, Los Angeles
Efficient natural gradient method for large-scale optimization problems 
We propose an efficient numerical method for computing natural gradient descent directions with respect to a generic metric in the state space. Our technique relies on representing the natural gradient direction as a solution to a standard least-squares problem. Hence, instead of calculating, storing, or inverting the information matrix directly, we apply efficient methods from numerical linear algebra to solve this least-squares problem. We treat both scenarios where the derivative of the state variable with respect to the parameter is either explicitly known or implicitly given through constraints. We apply the QR decomposition to solve the least-squares problem in the former case and utilize the adjoint-state method to compute the natural gradient descent direction in the latter case.

 

*********************************************************************************************************************************************************************

TrAC Seminar Series Fall 2021 

Meeting times: Thursday(s), 12:00--1:00pm (CST) 
Talk: 12:00 --12:50:00; Q&A: -- 12:50--1:00pm
Zoom  link:  
Contact: Hailiang Liu  hliu@iastate.edu, Soumik Sarkar,  Baskar Ganapathysubramanian    

The Translational AI Center for Research and Education (TrAC) is a newly established  ISU center,  a forum for integration of the activities and skill-sets of multiple ISU investigators with external collaborators  to accelerate knowledge transfer from fundamental scientific advances in AI  to industrial applications. The TrAC seminar meets Thursdays at the lunch time, with a lead speaker, raising issues and exchanging ideas on topics of current interest in the area of Translational AI and related advances.  The format consists of a lead presentation of about 50 minutes, followed by questions and  discussions.

10/14  
Adarsh Krishnamurthy  flyer
Mechanical engineering department, Iowa State University 
Title: Data-Driven Computational Modeling for Cardiovascular Mechanics

10/21
James E. Koltes  flyer
Department of Animal Science, Iowa State University 
Title: Development of Genetics, Genomics and Phenomics tools to enhance dairy cattle sustainability

10/28 
Iddo Friedberg flyer 
Department of Veterinary Microbiology and Preventive Medicine,  Iowa State University 
Title: What can an AI competition do for you?

11/04
Tichakorn (Nok)  Wongpiromsarn   flyer 
Department of Computer Science,  Iowa State University
Title:  Establishing correctness of learning-enabled autonomous systems
 

11/11
Julie Dickerson  flyer 
Electrical and Computer Engineering, Iowa State University
Title: Machine Learning for Biological Networks
 

11/18 
Aditya Ramamoorthy  flyer 
Electrical and Computer Engineering, Iowa State University
Title: Straggler Mitigation in Large Scale Distributed Matrix Computation


Talk info:

November 18 at 12:00 noonhttps://iastate.zoom.us/s/9378976918

Straggler Mitigation in Large Scale Distributed Matrix Computation
Aditya Ramamoorthy 
Electrical and Computer Engineering, Iowa State University

Abstract: High dimensional matrix computations are a key component of various algorithms within machine learning and scientific computing. Such computations are often deployed on large scale distributed computing clusters. The wide spread usage of these clusters presents several advantages over traditional computing paradigms. However, they also present new challenges, e.g., such clusters are well known to suffer from the problem of “stragglers (slow or failed nodes in the system ) which can end up dominating the overall job execution time.

In this talk we shall overview recent information-theoretic ideas inmitigating the effect of stragglers in distributed matrix computation. At a top-level these ideas allow the recovery of the desired result as long as any k-out-of-n worker nodes (where k<n ) complete their assigned tasks. The talk will highlight several open issues within this broad area and our recent work on these topics. These include, dealing with partial worker node computations (slow vs. failed nodes), sparse input matrices and ensuring numerical stability of the recovered result.

Short Bio:  Aditya Ramamoorthy is a Professor of Electrical and Computer Engineering and (by courtesy) of Mathematics at Iowa State University. He received his B. Tech. degree in Electrical Engineering from the Indian Institute of Technology, Delhi and the M.S. and Ph.D. degrees from the University of California, Los Angeles (UCLA). His research interests are in the areas of classical/quantum information theory and coding techniques with applications to distributed computation, content distribution networks and machine learning.

Dr. Ramamoorthy served as an editor for the IEEE Transactions on Information Theory from 2016--2019 and the IEEE Transactions on Communications from 2011--2015. He is the recipient of the 2020 Mid-Career Achievement in Research Award, the 2019 Boast-Nilsson Educational Impact Award and the 2012 Early Career Engineering Faculty Research Award from Iowa State University, the 2012 NSF CAREER award, and the Harpole-Pentair professorship in 2009 and 2010.
 

November 11 at 12:00 noonhttps://iastate.zoom.us/s/9378976918

Machine Learning for Biological Networks
Julie Dickerson 
Electrical and Computer Engineering, Iowa State University

Abstract: Biological systems are complex networks of genetic interactions and metabolism. Sensing technologies have allowed the external measurement of transcription levels, proteins, and metabolites. The challenge now is to use the measurements to help untangle the interactions in cells and organisms.

Bio: Julie Dickerson is  the David C  Nicholas Professor in Electrical and Computer Engineering. Her research program focuses on the application of data science to bioinformatics; this has led to successful collaborations with faculty across ISU. She has played a key role in the Bioinformatics and Computational Biology (BCB) Program as a past chair and as a core curriculum developer. Her development of the core systems biology course for all BCB students has enabled students to learn the basics of network science and data science as applied to biological systems. She has also served as an NSF Program Officer in the Advances for Biological Informatics in the BIO directorate.

 

November 4 at 12:00 noonhttps://iastate.zoom.us/s/9378976918

Establishing correctness of learning-enabled autonomous systems
Tichakorn (Nok)  Wongpiromsarn 
Department of Computer Science,  Iowa State University

Abstract: Autonomous systems are subject to multiple regulatory requirements due to their safety criticalnature. In general, it may not be feasible to guarantee the satisfaction of all requirements under all conditions. In such situations, the system needs to decide how to prioritize among them. Two main factors complicate this decision. First, the priorities among the conflicting requirements may not be fully established. Second, the decision needs to be made under uncertainties arising from both the learning-based components within the system and the unstructured, unpredictable, and non-cooperating nature of the environments.Therefore, establishing the correctness of autonomous systems requires a specification language that captures the unequal importance of the requirements, quantifies the violation of each requirement, and incorporates uncertainties faced by the systems. In this talk, I will discuss our early effort to partially address this problem and the remaining challenges.

Short Bio
Tichakorn (Nok) Wongpiromsarn received the B.S.degree in Mechanical Engineering from Cornell University in 2005 and the M.S. and Ph.D. degrees in Mechanical Engineering from California Institute of Technology in 2006 and 2010, respectively. She is currently an assistant professor in the Department of Computer Science at Iowa State University. Her research spans several areas of computer science, control, and optimization, including formal methods, motion planning, situational reasoning, hybrid systems, and distributed control systems. Most of her work draws inspiration from practical applications, especially in autonomy, robotics, and transportation. A significant portion of her career has been devoted to the development of autonomous vehicles, both in academia and industry settings. In particular, she was a principal research scientist and led the planning team at nuTonomy (now Motional), where her work focused on planning, decision making, control, behavior specification,and validation of autonomous vehicles.

October 28 at 12:00 noon https://iastate.zoom.us/s/9378976918

What can an AI competition do for you?
Iddo Friedberg
Department of Veterinary Microbiology and Preventive Medicine,  Iowa State University 

Abstract: In an era of high throughput biology, we increasingly rely on AI for classification of protein structure prediction to genome annotation. Moreover, AI methods are making their way into biomedicine and agriculture for diagnosing diseases and improving productivity. Yet how reliable are such methods? One way to both critique and improve AI methods is by competitions. I will discuss AI competitions in the biological sciences, with an emphasis on the latest results in protein structure and function prediction, as well as success stories and cautionary tales.

Bio: Iddo Friedberg is an Associate Professor in the department of Veterinary Microbiology and Preventive Medicine. He received his PhD from the Hebrew University of Jerusalem, and his postdoctoral training at the Sanford Prebys Burnham Medical Discovery Institute in La Jolla, California. He joined Iowa State university in 2015. His lab is applying computational methods to study metagenomics, protein sequence-structure-function connections, antimicrobial resistance, genome structure evolution, and CRISPR site prediction. He is a member of the Board of Directors of the International Society for Computational Biology and an organizer of the Critical Assessment of protein Function Annotation group, a consortium of 60 labs that is working to improve methods for protein function prediction

October 21 at 12:00 noon https://iastate.zoom.us/s/9378976918

Development of Genetics, Genomics and Phenomics tools to enhance dairy cattle sustainability
James E Koltes
Department of Animal Science, Iowa State University 

Abstract: Production of milk and dairy products from cattle are critical to providing essential nutrients to a growing population. Considerable effort has been dedicated to collecting and cataloguing a wealth of phenotypes in dairy cattle to improve production efficiency. Genetic selection in dairy cattle has been the major driver of the improved efficiency. However, some phenotypes important to understanding efficiency are expensive, time consuming and laborious to collect. For these traits, new cost-effective phenotypes need to be discovered to improve upon current selection strategies. The focus of my groups research program has been to develop new genetic, molecular and sensor-based tools to help improve feed efficiency, animal health, and correlated factors that can be used to improve dairy cattle feed efficiency, welfare and resilience to illness. Our research has identified novel sensor measures as new information sources for use in understanding variability in feed intake in dairy cattle. Current and future research is focusing on the use of molecular phenotypes such as metabolites from blood and milk as well as image data to improve our understanding of the genetics of efficiency in cattle. The long-term goal of our research is to identify genetic variants and mechanisms responsible for variability in health and efficiency related traits for use in selection as well novel phenotypes to monitor health and feed intake.

Bio:  Dr. James Koltes is an Assistant Professor in the Department of Animal Science within the Animal Breeding and Genetics group at Iowa State University. Dr. Koltes received his BS in Dairy Science and Genetics from the University of Wisconsin-Madison and PhD from Iowa State University in Genetics. His research at focuses on the use of new tools such as sensors and biomarkers in the genetic improvement of feed efficiency and health in dairy cattle. He also works on development of computational tools and resources to advance the application of genomics in livestock breeding. He also serves as the co-coordinator for the NRSP8 USDA multistate Bioinformatics program which oversees tools development, data sharing, database development and training programs for livestock genomics researchers in the United States

October 14 at 12:00 noon https://iastate.zoom.us/s/9378976918

Data-Driven Computational Modeling for Cardiovascular Mechanics
Adarsh Krishnamurthy 
Mechanical engineering department, Iowa State University 

Abstract: Cardiovascular diseases, such as heart failure, are one of the leading causes of death in the U.S. and pose a severe burden to the healthcare system. Most current treatments for cardiovascular diseases are based on rough estimates of outcomes from the results of clinical trials, which might not apply to individual patients due to patient-specific variations. Data-driven computational models of the cardiovascular system, developed from patient-specific clinical data, can help refine the diagnosis and personalize the treatment. In this talk, I will present recent advances in computational modeling that enable the simulation of a full four-chamber cardiac model. We have developed tools to generate a patient-specific high-order four-chamber cardiac mesh and use isogeometric analysis to simulate an entire cardiac cycle. The second part of the talk will focus on novel machine-learning algorithms to optimize the design of bioprosthetic heart valves. Machine-learning tools can significantly accelerate biomechanics simulations, leading to the development of a high-fidelity surrogate model. This surrogate model can then be used for optimizing the geometry of bioprosthetic valves, leading to patient-specific valves with better fit and performance, reducing the need for premature valve replacements. Finally, I will present some recent results in scientific machine learning in solving parametric partial differential equations (PDEs). The tools and methods developed in this research will help improve patient care and treatment outcomes, ultimately benefiting society with improved healthcare.

Bio: Adarsh Krishnamurthy is an associate professor in the mechanical engineering department at Iowa State University, where he currently leads the Integrated Design and Engineering Analysis (IDEA) lab. He was a post-doctoral researcher at UC San Diego and received his Ph.D. from UC Berkeley before this. He was the recipient of the NSF CAREER award in 2018 for developing GPU-accelerated tools for patient-specific cardiac modeling. His research interests include computer-aided design (CAD), GPU and parallel algorithms, machine learning, biomechanics, patient-specific heart modeling, solid mechanics, and computational geometry. His lab is currently funded by the NSF, ARPA-E, NASA, NIH, and the ONR.