Bring High Performance Computing to Everyone in College of Engineering at Texas A&M University!
Dr. Jian Tao joined the Department of Visualization in the College of Architecture on Sept 1st, 2021. His latest profile can be found at
Although he will continue working with engineering researchers and students, this website won’t be updated anymore. Please feel free to contact Jian at jtao@tamu.edu if you have any questions. Thanks!
Please join Texas A&M High Performance Research Computing and the MathWorks for a complimentary hands-on workshop for TAMU students and faculty.
Date: Friday, March 5, 2021
Time: 10:00 p.m.-1:00 p.m. Central Time (US and Canada)
Venue: WebEx Event
During this 3-hour self-paced, hands-on workshop, you will be introduced to parallel and distributed computing in MATLAB for speeding up your application and offloading work. By working through common scenarios and workflows, you will gain an understanding of the parallel constructs in MATLAB, their capabilities, and some of the typical issues that arise when using them.
Highlights include:
https://sciml.tamids.tamu.edu/
Scientific Machine Learning Lab
TAMIDS has created a new program to collaborate with faculty teams in establishing Labs in emerging areas of Data Science. The mission of each Lab will be to develop knowledge, resources, and community around a thematic area of Data Science, encompassing research, education, and outreach. TAMIDS will support each Lab through a combination of seed funding for new research, effort from TAMIDS personnel, preferred access to existing TAMIDS programs, and organizational and logistical support. TAMIDS is piloting this program through its Scientific Machine Learning Lab (SciML Lab).
Scientific Machine Learning
Scientific Machine Learning (SciML) is a rapidly developing area that aims to revolutionize the practice of Science and Engineering, by bringing together the fields of Machine Learning and Scientific Computation. Typical data-driven Machine Learning methodologies do not incorporate physical understanding of the problem domain. Furthermore, in many scientific domains high-fidelity data are expensive or time-consuming to obtain. Physics-aware SciML addresses both of these problems by introducing regularizing constraints obtained from physical laws, allowing prediction of future performance of complex multiscale, multiphysics systems using sparse, low-fidelity, and heterogeneous data. Unlike traditional black-box Machine Learning methods, SciML aims to deliver interpretable models, leading to improved verification and validation in mission-critical applications.
SciML Lab Goals
There is a great unexplored potential for research and education in Scientific Machine Learning, with many opportunities for high-impact, highly-visible research and extra-mural funding over the next decade. Texas A&M University, with its large research enterprise in engineering, science, and high-performance computing, is uniquely positioned to emerge as one of the national leaders in this area. The purpose of SciML Lab is to help to make this a reality. The lab will foster community building efforts by means of workshops and a seminar series, will develop new educational resources for SciML, including tutorials, hands-on short courses, and a for-credit courses with stacked sections at the undergraduate and graduate levels, all of which will be aimed at audiences drawn from multiple disciplines within Science, Engineering, Geosciences and Agriculture and also promote knowledge of SciML in further disciplines that present opportunities for impact.
The Master of Science in Data Science degree is an on-campus interdisciplinary program offered by the Departments of Computer Science and Engineering, Electrical and Computer Engineering, Mathematics, and Statistics within the University’s Colleges of Engineering and Science, and administered jointly with the Texas A&M Institute of Data Science.
Application Deadline for Fall 2021: March 31
Eligibility Requirements: Applicants will need to have certain course work prior to enrolling in the program:
Math: calculus and linear algebra
Statistics: college level introduction to statistics
Some programming experience in at least one of languages: R, Python, C++.
Applicants with a bachelor’s degree in mathematics, statistics, computer science, electrical engineering, industrial engineering or similar fields should have the sufficient background.
HPRC and TAMIDS Workshop: Data Visualization and Geospatial Analysis With R
Thursday, November 12, 1:00 p.m. to 5:15 p.m.
More details about the workshop can be found at the registration page at
http://bit.ly/R_Viz_Workshop_2020
October 27, 2020
8:30 am – 1:00 pm
Online via zoom
https://tamids.tamu.edu/2020/10/07/wkshp-sciml/
Scientific Machine Learning (SciML) is a rapidly developing area that aims to revolutionize the practice of Science and Engineering, by bringing together the fields of Machine Learning and Scientific Computation. Typical data-driven Machine Learning methodologies do not incorporate physical understanding of the problem domain. Furthermore, in many scientific domains high-fidelity data are expensive or time-consuming to obtain. Physics-aware SciML addresses both of these problems by introducing regularizing constraints obtained from physical laws, allowing prediction of future performance of complex multiscale, multiphysics systems using sparse, low-fidelity, and heterogeneous data. Unlike traditional black-box Machine Learning methods, SciML aims to deliver interpretable models, leading to improved verification and validation in mission-critical applications.
This informal workshop is intended to help identify and bring together members of the academic community at Texas A&M who are interested in studying SciML. Areas of interest include: physics-informed deep neural networks, data-driven model discovery through large-scale simulation, ML-guided acceleration of numerical simulations; ML-guided automation of data acquisition and decision-support for complex systems; incorporating approaches such as: multifidelity surrogate modeling, uncertainty quantification, Bayesian inference; and computational frameworks, systems, and methods for SciML. While recent advances in SciML have been driven primarily by applications in Engineering, Physical Sciences and Synthetic Biology, the workshop encourages discussion of potential applications in other fields that may integrate domain knowledge and models with machine learning. Through the exchange of ideas and results we hope to foster coherent research efforts that can lead to new scientific advances and enhance competitiveness for external funding.
The workshop is open to all members of Texas A&M via zoom meeting ID 945 9928 7403 passcode 526915 (TAMU authentication required). No registration is required to attend. The workshop will comprise:
Currently confirmed speakers are listed below. The full program will be announced during the week prior to the workshop.
Members of Texas A&M who wish to contribute a short slide presentation to appear on the workshop website (up to 5 slides, pdf format) conformant to the scope of the workshop should send these by email to Ms. Jennifer South jsouth@tamu.edu with the subject line “SciML Workshop Contributed Presentation”, preferably by end October 20, 2020.
Ulisses Braga-Neto (ECE), Nick Duffield (TAMIDS / ECE), Jian Tao (TEES / TAMIDS / HPRC / ECE), with thanks to Narasimha Reddy (ECE / TEES) for fostering discussion on Scientific Machine Learning.
The Texas A&M Institute of Data Science (TAMIDS), the Texas A&M Engineering Experiment Station (TEES), Texas A&M High Performance Research Computing (HPRC)
Ms. Jennifer South, TAMIDS, jsouth@tamu.edu
TAMIDS Tech Talk on Tuesday, November 10th , by Vinit Sehgal, Ph.D. student in the Water Management and Hydrological Science Program at Texas A&M University.
More about the talk can be found at https://tamids.tamu.edu/2020/10/17/tamids-tech-talk-vinit-sehgal-large-scale-geospatial-analysis-with-r/
Mathworks and Texas A&M High Performance Research Computing are offering two introductory online sessions about data analysis and parallel computing with Matlab. These sessions are offered free of charge and are available to anyone at Texas A&M. Registration is required.
Wed October 28 | 1PM-3PM | WebEx session
Parallel Computing with MATLAB
Prerequisites: basic MATLAB knowledge
This session will discuss a number of features of the MATLAB Parallel Toolbox to speedup computations. This session will use live demonstrations and examples.
Wed October 28 | 3PM-5PM | WebEx session
Data Analysis with Matlab
Prerequisites: None
This session will use live demonstrations and examples to discuss how Matlab can be used to perform data analysis.
To see the complete catalog of short courses offered by High Performance Research Computing, please visit hprc.tamu.edu/training.
Speaker: Kailai Xu, Ph.D. Student at Stanford University
Date: Tuesday, October 6, 2020
Time: 3:00 – 4:00PM US Central Time
Zoom Link: https://tinyurl.com/yy8ldtfh
Abstract
ADCME is a novel computational framework to solve inverse problems involving physical simulations and deep neural networks (DNNs). By describing physical laws with partial differential equations (PDEs) and substituting unknown components with DNNs, we preserve the physics to the largest extent while leveraging DNNs for data driven modeling. To train the DNNs within a physical system, ADCME expresses both numerical simulations (e.g., finite element method) and DNNs as computational graphs and calculates the gradients using reverse-mode automatic differentiation. We have built a system of re-usable and flexible numerical simulation operators that support gradient-backpropagation for many engineering applications, such as seismic inversion, constitutive modelin g, Navier-Stokes equations, etc. ADCME also provides a computational model for conducting large-scale inverse modeling using MPI, and has been deployed across thousands of cores. The ADCME software is open -sourced and available at https://github.com/kailaix/ADCME.jl.
Biography
Kailai Xu is a Ph.D. student in computational mathematics at Stanford. His current research interest centers on machine learning for inverse problems in computational engineering. He has developed the open-source software ADCME.jl in Julia and C++ for high-performance inverse modeling using automatic differentiation. Specifically, he has developed novel physics-based machine learning algorithms and software packages based on ADCME.jl for solving inverse problems in stochastic processes, solid mechanics, geophysics and fluid dynamics. One highlight of his research is combining neural networks with numerical solvers for PDEs, which enables data-driven modeling with physics knowledge.
Weekly Online Bring-Your-Own-Data (BYOD) Workshops
Aug 26 – Dec 16, 2020
10:00 am – 11:30 am
Wednesdays online meeting via Zoom
The Texas A&M Institute of Data Science (TAMIDS), Texas A&M High Performance Research Computing (HPRC), and Texas Engineering Experiment Station invite you to join our Bring-Your-Own-Data (BYOD) workshop series!
These are one-on-one consultancy sessions with a TAMIDS Data Scientist who can help with formulating approaches to your Data Science research project, and assist with code development to take advantage of the latest data analytics methods and high-performance computing facilities.
This is a FREE service offered to all researchers at Texas A&M.
Sign up at Eventbrite (https://tinyurl.com/y2cjkbuz)