Computational Engineering Initiative at FAMU-FSU College of Engineering...
Prepared by Kourosh Shoele†
POC: kshoele@eng.famu.fsu.edu, +1-850-645-0143
Participants
Unnikrishnan Sasidharan Nair†, Neda Yaghoobian†, Christian Hubiki†,
Leo Liu¶,
Sungmoon Jung‡, Pedro Fernandez-Caban‡, Ebrahim Ahmadisharaf‡,
Bayaner Arigong↑, Victor DeBrunner↑, Rodney Roberts↑,
Hui Wang‖, Lichun Li‖, Yanshuo Sun‖, Hui Wang‖
†Mechanical Engineering
¶Chemical & Biomedical Engineering
‡Civil & Environmental Engineering
↑Electrical & Computer Engineering
‖Industrial & Manufacturing Engineering
Prepared for the Dean Suvranu De;
Compiled: Jan 15, 2024
Introduction...
The convergence of computational engineering and machine learning is transforming scientific and industrial landscapes across various engineering fields. Industries are strategically leveraging these technologies, particularly for tackling intricate optimization challenges in design and manufacturing. Machine learning, considered as data-driven optimization, excels in addressing high-dimensional, non-convex, and constrained multi-objective optimization problems, demonstrating improvement with larger datasets computed through advanced computational techniques. This underscores the imperative to educate new engineers proficient in potent computational techniques and interpretable, generalizable, explainable, and certifiable machine learning techniques. These skills are crucial for advancing safety-critical applications in aerospace, biomedical, electrical, and mechanical systems.
Observing our progress in computational techniques and machine learning, the anticipation is towards the development and progression of Digital Twin—a virtual replica integrating machine learning and multiscale modeling to dynamically adapt in real-time. Future engineers need to be well-equipped to explore new potential applications in healthcare, engineering, and education by integrating population and case-specific data. This is only one potential benefit of providing education in computational techniques and machine learning. A quick search reveals the rapidly growing job market for computational modeling and machine learning, driven by increasing demand across various industries. This growth is rooted in several factors:
- Rising complexity of engineering systems: Modern engineering challenges involve intricate behaviors at multiple scales, demanding advanced modeling techniques like multiscale modeling.
- Data explosion and the need for data-driven insights: The vast amount of data generated in various sectors requires machine learning for analysis and prediction, often in conjunction with simulations.
- Focus on efficiency and innovation: Multiscale modeling and machine learning can expedite design and optimization processes, leading to faster development cycles and improved product performance.
The FAMU-FSU College of Engineering can play a pivotal role in training highly qualified engineers who can enter the workforce in diverse sectors:
- Research and development: Roles in academia, national labs, and private companies focusing on developing and applying multiscale modeling and machine learning techniques to solve specific engineering problems.
- Product design and development: Crucial for designing and optimizing products in various industries such as aerospace, automotive, energy, materials, and healthcare.
- Data science and analytics: Professionals with skills in both machine learning and domain knowledge are needed to analyze data generated from simulations and experiments, extract meaningful insights, and inform decision-making.
- Computational modeling and simulation: Jobs involve developing and running simulations using multiscale models and incorporating machine learning for enhanced accuracy and efficiency.
The future salaries for professionals with expertise in multiscale modeling and machine learning are expected to be competitive, exceeding the average salaries for conventional engineers in specific industries.
Moreover, the job market for individuals skilled in multiscale modeling and machine learning is robust and diverse, spanning various industries and sectors. Here are some areas where professionals with expertise in multiscale modeling and machine learning are in demand:
- Biomedical Research and Healthcare:
- Developing computational models for biological systems and physiological conditions.
- Analyzing medical data for diagnostics and treatment optimization.
- Pharmaceuticals and Drug Discovery:
- Predicting molecular interactions and patient-specific drug delivery modeling.
- Applying machine learning in drug discovery biochemical processes.
- Material Science and Engineering:
- Designing new materials with enhanced properties.
- Predicting material behaviors under different conditions.
- Climate and Environmental Science:
- Modeling climate systems and environmental processes.
- Analyzing large-scale environmental data using new techniques
- Aerospace and Engineering:
- Simulating and optimizing complex engineering systems.
- Applying machine learning for predictive maintenance and optimization.
- Energy Sector:
- Optimizing energy systems and predicting energy-related phenomena.
- Applying machine learning for energy consumption optimization.
- Robotic and Automotive Industries:
- Simulating and optimizing vehicle/robot performance.
- Using machine learning for autonomous technologies.
- Research and Academia:
- Conducting research in computational science and engineering at national labs.
- Advancing the education of multiscale modeling and machine learning.
- Data Science and Analytics:
- Analyzing large datasets using machine learning techniques.
- Extracting insights and patterns from complex data structures.
Currently, there is a trend of initiating similar programs in various schools, and the CoE can explore this research theme to attract talented researchers from these programs. Many colleagues and their groups are engaged in computing; however, limited training in numerical mathematics, algorithms, and software engineering hinders their progress in dynamic research fields, especially with recent advancements driven by superior algorithms rather than hardware.
On a larger scale, a fundamentally new paradigm is emerging for creating models and conducting engineering designs and scientific discovery. The argument is made that we are entering the era of the multicomputational paradigm, surpassing previous modeling paradigms such as structural, mathematical, and computational paradigms. Examining the historical evolution of science reveals these three major paradigms, each contributing significantly to progress. The structural paradigm was an early method where the world's elements were considered constructed from simple-to-describe elements, using logical reasoning for understanding complex systems. The mathematical paradigm, prevalent for three centuries, described things in the world using mathematical equations. It was succeeded by the computational paradigm, emphasizing the use of simple programs for modeling natural phenomena. However, the emergence of new and more complex problems necessitates a broader paradigm—the multicomputational paradigm. Extending beyond physics and engineering, this paradigm provides a versatile methodology for theoretical science, emphasizing the simultaneous use of multiple computational processes in modeling complex systems.This necessitates new academic engineering education that differs from conventional techniques. In particular, the student should be trained to leverage new computational techniques to better characterize complex systems. Future education should enable students to go beyond single-domain skills and equip them with sufficient knowledge to tackle multidisciplinary engineering problems.
Proposal...
We propose to establish a computational and data-enabled engineering FAMU-FSU College of Engineering (COE) aiming to establish and administer a graduate program leading to the Designated Emphasis in High performance Computation and Data-enabled Engineering (HCDE). The HCDE is applicable to all existing programs at the FAMU-FSU COE and currently, the College of Engineering has invested in disciplines relevant to the HCDE program, including solid mechanics, materials science and fluid and thermal systems. Our step-by-step approach involves key aspects of the initiative, including the selection and role of the Executive Committee and the definition of program standards covering admissions, curriculum, and degree requirements.
The program aims to provide students with a well-rounded education in computational engineering and cyber infrastructure, fostering multidisciplinary knowledge. The curriculum includes foundational and applied courses covering modeling and computation, analytics, algorithms, high-performance computing, data management, analysis, visualization, software, and multidisciplinary collaborations. To earn the certificate, students must choose a set number of courses from a core selection specific to the program. Furthermore, they are required to select elective courses from various domains, including analytics, computational methods in engineering, medical/health, control, and computer visualization.
Questions and Answers:...
As the discussion on initiating a computational engineering initiative in FAMU-FSU CoE commenced, the Dean and COE administrators raised several questions. In response, multiple meetings were conducted, inviting COE-affiliated faculties with expertise in various aspects of computational engineering. After several group meetings and individual communications, we could identify several important aspects. The following questions and answers serve as a summary derived from the input of faculties who participated in these meetings.
I. Certification Exploration:...
1. Is it viable to launch a certification in computational engineering?
The responses overwhelmingly support the viability of launching a certification in computational engineering. Participants think about the need for at least four courses for the certificate (2 core courses and 2 specific classes from a dedicated list in the area of student’s engineering domain) and express the importance of making computation visible in the school. There is a recognition that many students are unaware of the potential use of computation in research, particularly given the prevalence of experimentalists in the school. The multidisciplinary nature of the certification is acknowledged, with a cautious note on the inclusion of physics-based simulation and data sciences. The responses stress the significance of a carefully designed course that covers the contents comprehensively, including backgrounds in solid mechanics, dynamics, fluid dynamics, FEA, statistical machine learning, and basic coding skills for machine learning. Overall, the consensus is positive, with confidence that with a thoughtful selection of relevant coursework, the launch of the certification is indeed possible. Additionally, the certification can be incorporated into the agendas of both universities and utilized to attract substantial federal grants. This initiative has the potential to lay the foundation for an HPC center at FAMU, situated in the innovation park. Such a center can serve as a distinctive infrastructure, emphasizing the innovative training approach in high-performance computing and data-enabled engineering.
2. Which skills and knowledge areas should be prioritized for this certification?
The prioritized skills and knowledge areas for the certification in computational engineering encompass a broad spectrum. Respondents emphasize the importance of knowing the bases of computation and data-enabled techniques. Other potential skills are related to simulation and optimization, data science, and programming in Python/Matlab/C. The need for a clear distinction from computer scientists is necessary, suggesting a balanced approach that combines computer science-like skills, such as machine learning, with specialized software skills in areas like CFD, FEM, and optimization. High-performance computing, parallel computing, distributed computing, and cloud computing are also crucial areas, with a note on the potential reluctance of some of engineering disciplines if the core course is primarily focused on mechanics. This can be solved through careful planning of the program and by ensuring the courses offer a more general introduction to several topics of interest to all engineering domains.
We suggest the development of a distinctive program that integrates applied computational engineering with a close alignment between course content and the experimental/computational strengths of our research centers. Such dual training will attract nearby professional engineers to employ new techniques for their systems and will enable the transition of our current experimental facilities to consider data-driven approaches in their portfolio and explore digital-twin topics. We also highlight the importance of including theoretical and practical aspects of numerical analysis, applying courses in computational engineering, and leveraging HPC capabilities and machine learning tools for data analysis and simulation. Moreover, there was a consensus among all of us on the significance of establishing a strong foundation in fundamental math, applied math in engineering contexts, basics of scientific computing, and hands-on coding exercises for the certification. The program's goal should be to provide an education encompassing a balanced skill set, incorporating both theoretical understanding and practical application in the field of computational engineering.
3. Can we lay out a concrete plan and set timelines? This may require the committee members to work with the department chairs.
Our discussion showed varying levels of commitment to laying out a concrete plan and setting timelines for the certification in computational engineering. While all of us agree about the plan in general, more discussion is needed for specific suggestions and timelines. We suggest creating an executive committee panel to explore the plan and practical timeline and to engage faculties in different departments. We think of spending 1-2 semesters through interactive discussions to create a course plan, emphasizing exploring existing courses before considering new courses. We think that an initial draft curriculum can be compiled in spring 2024, with department buy-in targeted by the end of Summer. There is a consensus that the feasibility of the plan is acknowledged, with an emphasis on careful consideration, especially for implementation by Fall 2024.
II. Visibility and Impact:...
1. Are there specific plans or strategies in mind to boost the profile and visibility of the Computational Engineering program?
Our discussion suggests making the certification available online, emphasizing accessibility. Also, certain unique aspects such as the practical digital twin framework, can be used to enhance visibility. Perhaps other aspects could be the determination of courses, designation of instructors, and collaboration with FSU ODL and FAMU OIT for online program development. We think it is necessary to publicize the certification early, targeting second-year undergraduates and encouraging motivated students to pursue the computational track. The feasibility of these initiatives is acknowledged, emphasizing inclusivity by stating that the program shouldn’t be limited to students performing computational research but can also benefit those engaged in physical experimentation, enhancing efficiency in data analysis. Moreover, through the inclusion of practical approaches in domain-specific topics we can attract nearby industries to enroll in the program.
III. Graduate Recruitment:...
1. What are some ideas to create a robust pipeline of computational engineering graduate students?
One prominent suggestion was to provide one/two-year scholarships annually, specifically targeting high-quality students interested in computational research. Collaboration with computing departments for recruitment is also proposed, emphasizing strategic partnerships. The development of an online program with asynchronous instruction was another suggestion to attract part-time or non-degree-seeking graduate students who are currently employed, with encouragement for current graduate students to participate. The use of a dedicated webpage and effective advertisement is seen as a starting point, followed by the pursuit of external grants. Additionally, creating motivation through lower-level classes, highlighting the program’s benefits for both computational researchers and experimentalists, and actively showcasing FSU’s computational capabilities through media and outreach initiatives are recommended strategies.
- What's our approach to fostering a multidisciplinary research and educational experience for our graduate students?
We have identified several approaches to promote multidisciplinary research. One suggestion involves the formation of focus teams, consisting of faculty and scholars from diverse fields, collaborating on joint research and capstone projects for students. To encourage actual collaboration and mentoring, another proposal is to pursue "seed funding" specifically for student recruitment, with a requirement for multi-departmental/topic collaboration to enhance overall funding prospects.
The concept of joint courses and seminars is presented as a method to create a more multidisciplinary environment. The program can explore collaborations with national labs such as Oak Ridge and NREL, as well as public and private research groups like Mayo Clinic, Eglen, etc. It is recommended to provide practical examples of societal challenges that necessitate multidisciplinary collaboration, highlighting the program's impact through the facilitation of computational skills. Additionally, proposing numerical methods as a core course in undergraduate engineering is suggested as a foundational step in advancing multidisciplinary education.
3. How can we ensure students get practical experience with advanced computing?
Our first suggestions include introducing a Research Computing Center (RCC) resources in computational engineering courses, emphasizing its integration into individual faculties. Comprehensive projects within courses are proposed, challenging Research Assistants (RAs) to apply learned skills to solve research problems in their disciplinary projects. We unanimously agreed that the requirement of a selling point for computational engineering compared to computer science, by promoting a balanced skill set with a focus on engineering and more practical software skills. Developing a course on advanced computing topics, jointly determined by interested faculty from multiple engineering departments, is recommended. This course should be tied to high-performance computing facilities to maximize its impact. Additionally, higher-level courses incorporating nontrivial implementations, focusing on real-world practical problems, and demonstrating the cost-effectiveness and efficiency of advanced computing tools could be other effective approaches.
IV. Computational Needs Assessment:...
1. Can we identify the computational requirements across our engineering disciplines?
The responses across engineering disciplines vary. While some were able to identify department-level requirements, while others provided no specific information. However, the group agreed that multi-core CPUs with large memory for large-scale optimization and GPUs (specifically Nvidia A6000 or Mac M2 Ultra) for parallelization in handling large computations and processing images/videos are important requirements. In the context of Civil and Environmental Engineering (CEE), it is noted that the previous engineering cluster provided reasonable resources which enables new research. This is highlighted by the leverage of High-Performance Computing (HPC) resources and parallel processing for computational fluid dynamics simulation, structural optimization, and the development of advanced machine learning-based models in predicting wind performance of civil structures. We envision two specific hardware requirements:
1. A fast connection between our centers and a central high-performance computing center. With meaningful support from FAMU, we can consider establishing a new center that integrates mid-size infrastructure located in CoE for both data analysis and high-performance computing. We envision putting together practical Jupiter notebooks and step-by-step tutorials to enable practical education of the students participating in the program. The exact assessment of the future requirement is highly dependent of the CoE expansion and research policies.
2. Enabling our centers with fast (real-time) sensors and actuators as well as memory devices that enable integration of computational techniques into the current research thrusts in different centers. This can be done initially for targeted benchtop experiments by faculties involved in computational and experimental research approaches.
2. What challenges are researchers encountering in these domains? How can we turn these challenges into opportunities for creating a focal strength in computational engineering?
Researchers in computational engineering face a range of challenges. Issues include computing resources, student recruitment, and the quality of students. A call for cultural change in how various research types are perceived within departments is noted. Many new approaches have relied on large-scale optimization. To enable new discovery, it is necessary to use high-performance computing algorithms to enhance efficiency. Similarly, addressing parallelization issues involves the efficient use of CPU/GPU resources and this will enable seamless integration of multiphysics simulation techniques and machine learning inferences. The complexity of algorithms can be tackled by introducing physics-informed learning methods, reducing model complexity while improving interpretability. Setting up and updating software is identified as a challenge, with a suggestion for staff assistance. Other challenges include insufficient computational High-Performance Computing (HPC) resources and slow student learning curves, with a recommendation to increase awareness of state-of-the-art computational capabilities at peer institutions and view Research Computing Center (RCC) as basic infrastructure rather than a paid service.
One the bigger scales, most of similar programs are following the high performance computing (HPC) and artificial intelligence/machine learning (AI/ML) as a separate research and education efforts. Yet, the opportunity is to integrate the current methods in HPC and AI/ML and create a new research paradigm that benefits from both domains. This would be inline with the current trend of funding availability by different agencies.
3. Do we have a defined annual budget for RCC? Should we look beyond RCC? Has the HiPerGator at UF been on our radar?
The initial CoE contribution to RCC was based on an annual budget of 50K, which was not consistently followed in the previous administration. Although the current RCC prices have been reduced to encourage usage, the model remains inefficient for sustainable research. The centers are suitable for traditional small to moderate-size computations, and acquiring both CPU/GPU from the center can be costly. Additionally, issues such as node and memory expiration disrupt faculty activities, and immediate external funding for resource repurchase may not always be available.
Responses to the question regarding the effectiveness of the Research Computing Center (RCC) and the consideration of alternatives reveal varied perspectives. Exploring options beyond RCC is seen as an opportunity to engage centers at both universities, attracting maximum support and enabling longer use of investments. The decision may depend on the required number of NCUs and storage space, with emphasis on RCC resources, which have been mostly sufficient for individual research use thus far. HiPerGator is currently dedicated to memory-intensive computations and AI/ML tasks, and its speedup for multiphysics problems may not be optimal. Exploring potential partnership avenues with HiPerGator is recommended for further consideration by our school.
V. Current Classes with data science or numerical aspects in Your Department...
1. Chemical & Biomedical Engineering...
1. BME3702 - Biocomputations
2. ECH5840 - Advanced Chemical Engineering Mathematics
2. Civil & Environmental Engineering...
1. CES 6116 | Finite Elements Methods
2. CCE 5510 | Computer Applications in Construction
3. EGN 5458 | Statistical Applications for Engineers
4. EGN 5480 | Metaheuristics and Hybrid Algorithms
3. Electrical & Computer Engineering...
1. EEL 5930 - Computational Intelligence
2. ENG2520 - Pattern Recognition and Machine Learning
3. EEL4450 - Modeling and Simulation of Semiconductor Devices
4. EEL 5205 - Computational Methods in Electrical Engineering
5. EEL 5452 - Modeling and Simulation
6. EEL 5930 - Computational Methods in Power Systems
4. Industrial & Manufacturing Engineering...
1. EGN 5444 - Big Data Analytics in Engineering
2. ENG2520 - Pattern Recognition and Machine Learning
3. ESI 5458 - Optimization on Networks
4. ESI 5685 - Introduction to Machine Learning
5. ESI 5243 - Engineering Data Analysis
6. ESI 5681 - Deep Learning in Practice
7. EML 5930 - Uncertainty Analysis
5. Mechanical Engineering...
1. EML4930(EGN 3434) - Numerical Methods for Engineers
2. EML4930/5930 - Computational Material Physics
3. EML5537 - Design using Finite Element Methods
4. EML5930 - Computational Fluid Dynamics for Incompressible Flows
5. EGM5611 - Continuum Mechanics
6. EML5930 – Model Reduction
7. ESI 5408 – Applied Optimization
8. EML 4930 - Computational Linear Algebra
9. EML5930 - Advanced Numerical Method
10. EML5930 - Fluid Structure Interaction
6. Potential Other Classes...
- DATA SCIENCE
* Intro Machine Learning- Intro Theoretical Statistics
- Design of Experiments
- Bayesian Statistics
- Big Data/Machine Learning
- APPLIED NUMERICAL MATHEMATICS
- Introduction to Complex Systems
* Method of Applied Math - Network Theory
- Applied Functional Analysis
- Bioinformatics
- Quantum Computing
- Introduction to Complex Systems
- HIGH PERFORMANCE COMPUTING
- Data-Oriented Computing for Engineers
- Introduction to Parallel and Distributed Processing
- Data Intensive Computing
- High Performance Computing for Engineers
- OTHERS
- Concepts in Multi-Scale and Multi-Physics Modeling
- Applied Probability and Statistics
- Computational Modeling and Visualization
- Computational Statistics
- Introduction to Computational Software Tools
Appendix...
1-Other examples...
Rutgers Web...
Summary...
Rutgers University offers a graduate certificate in Computational and Data-Enabled Science and Engineering (CDS&E) through its Discovery Informatics Institute and the Professional Science Master's Program. The program aims to provide a multidisciplinary experience to students, enhancing their understanding of science, engineering, and business through advancements in cyber infrastructure. The curriculum covers foundational and applied courses in modeling and computation, analytics, algorithms, high-performance computing, data management, analysis, visualization, software, and multidisciplinary collaborations.
To earn the certificate, students must choose two courses (6 credits) from a list that includes "Introduction to Cloud Computing and Big Data," "Introduction to Parallel and Distributed Computing," and "Parallel and Distributed Computing," among others. Additionally, students must select two elective courses (6 credits) from various domains such as analytics, computational methods in engineering and sciences, finance, social science, medical/health informatics, and computer visualization.
Furthermore, students are required to attend at least six colloquia in CDS&E, focusing on related topics listed on the Rutgers Discovery Informatics Institute website. This comprehensive program is open to all graduate students in the sciences and engineering, emphasizing practical skills and knowledge in the evolving field of computational and data-enabled science and engineering.
Certificate Program...
The graduate certificate in computational and data-enabled science and engineering (CDS&E) is a cross-disciplinary graduate program based in the Rutgers Discovery Informatics Institute (RDI2) and the Professional Science Master's Program (Master of Business and Science degree). The goal of the program is to provide the necessary structures, learning opportunities, and experiences, beyond the more traditional university curriculum, that are necessary to drive science, engineering, and business using advances in cyber infrastructure (CI). The program will provide an overlay on the existing curricular structures at Rutgers to give students a multidisciplinary experience, and will include foundational and applied courses spanning modeling and computation, analytics, algorithms, high-performance computing, data management and analysis, visualization, software, and multidisciplinary collaborations.
The certificate in CDS&E is open to all graduate students in the sciences and engineering. To receive the certificate, students must complete all the course requirements listed below.
2 courses from this list (6 credits):...
16:137:602 Introduction to Cloud Computing and Big Data (3cr)
This course introduces fundamental concepts, technologies and innovative applications of Cloud and Big Data systems like distributed systems, map reduce programming model, distributed file systems virtualization and cloud models, etc. Engineering aspects like bridging the gap between analytics and data-driven platforms, performance evaluation and benchmarking. Explore recent technological solutions and research in Cloud and Big data. Hands on experience with in Hadoop, HDFS and big data databases, SQL, noSQL and newSQL.
16:332:566 (S) Introduction to Parallel and Distributed Computing (3cr)
Systems, architectures, algorithms, programming models, languages, and software tools. Topics covered include parallelization and distribution models; parallel architectures; cluster and networked metacomputing systems; parallel/distributed programming; parallel/distributed algorithms, data-structures and programming methodologies; applications; and performance analysis. Programming assignments and a final project. Prerequisites: 16:332:563 and 564.
16:332:572 (S) Parallel and Distributed Computing (3cr)
Advanced topics in parallel computing including current and emerging architectures, programming models application development frameworks, runtime management, load balancing, and scheduling, as well as emerging areas such as autonomic computing, grid computing, pervasive computing, and sensor-based systems. Prerequisite: 16:332:566.
16:137:552 Python for Data Science (3cr)
This course covers the basics of Python and how it can be used for data science and computationally enabled science and engineering. A major project involving data is part of the course.
Please choose two elective courses (6cr)
- Analytics
- Computational methods in engineering (biomedical, chemical, civil, electrical, industrial, mechanical, etc.) - For example, computational solid and fluid mechanics, finite element analysis, computational aerodynamics, computer simulation of materials, etc.
- Computational methods in sciences (physics, chemistry, biology, etc.)
- Computational methods in finance
- Computational methods in social science
- Medical/health informatics
- Computer visualization
Students must attend at least six colloquia in CDS&E for the certificate. Colloquia in CDS&E-related topics will be listed on the Rutgers Discovery Informatics Institute website.
Brown Web...
Summary...
The Master of Science in Data-Enabled Computational Engineering and Science (DECES) program at Brown University's School of Engineering is positioned as a world leader in disciplines like solid mechanics, materials science, and fluid and thermal systems. Leveraging the strengths of their faculty in Engineering and Applied Mathematics, particularly in developing cutting-edge numerical methods and machine learning approaches, Brown offers a unique DECES ScM Program. This program is tailored for students aspiring to careers involving advanced modeling and simulation in engineering and physical sciences, and it is also suitable for working professionals relying on high-fidelity engineering simulations with data assimilation and machine learning expertise. The program aims to provide students with a profound understanding of the role of advanced simulation in industry and national laboratories, an appreciation for high-fidelity modeling and simulation in contemporary engineering design, and technical knowledge in foundational subjects of computational engineering. Additionally, graduates will possess the necessary skills, combined with machine learning expertise, to effectively conduct practical engineering-scale simulations.
Details...
The School of Engineering at Brown is a world leader in several disciplines relevant to the Master of Science in Data-Enabled Computational Engineering and Science (DECES) program, including solid mechanics, materials science and fluid and thermal systems. Brown has one of the highest-ranked Applied Mathematics programs in the nation. Many of our stellar faculty in Engineering and Applied Mathematics are working on developing state-of-the-art numerical methods and machine learning approaches, with applications that are of particular relevance to the DECES ScM program. As such, Brown is uniquely positioned to offer the DECES ScM Program.
The DECES ScM program was designed for students interested in pursuing careers that involve advanced modeling and simulation in engineering and physical sciences. It may also be of interest to working professionals whose success on the job depends on their ability to competently perform high-fidelity engineering simulations with data assimilation and machine learning expertise.
Upon completion of the program coursework the students will:
- Gain the understanding of a significant role that advanced simulation plays in industry and national laboratories
- Develop an appreciation for the power of high-fidelity modeling and simulation in contemporary engineering design
- Gain technical knowledge of the foundational subjects in computational engineering, including nonlinear finite element analysis and the integration of physics-based modeling and data science
- Develop the necessary technical skills combined with machine learning expertise to knowledgeably carry out practical engineering-scale simulations
Buffalo Web (Bauman et al., 2014)...
DATA SCIENCE...
CSE 574: Intro Machine Learning
STA 521: Intro Theoretical Statistics 1
STA 522: Intro Theoretical Statistics 2
STA 534: Design of Experiments
STA 567: Bayesian Statistics
MAE 701 Special Topics: Bayesian Methods in Engineering Applications
CSE 704 Seminar: Big Data
CSE 740 Seminar: Big Data/Machine Learning
APPLIED NUMERICAL MATHEMATICS...
MTH 537: Introduction to Numerical Analysis 1
MTH 538: Introduction to Numerical Analysis 2
MTH 539: Method of Applied Math 1
MTH 540: Method of Applied Math 2
MTH 550: Network Theory
MTH 555: Introduction to Complex Systems
MGF 636: Complex Fin Instruments
MAE 702 Seminar: Applied Functional Analysis
HIGH PERFORMANCE AND DATA INTENSIVE COMPUTING...
MTH 548: Data-Oriented Computing for Mathematicians
CSE 570: Introduction to Parallel and Distributed Processing
CSE 587: Data Intensive Computing
CDA 609: High Performance Computing 1
CDA 610: High Performance Computing 2
UC Berkely Web...
Summary...
The College of Computing, Data Science, and Society at UC Berkeley sponsors a Designated Emphasis in Computational and Data Science and Engineering (DE-CDSE). This interdisciplinary graduate minor focuses on developing curricula and programs to advance the use of numerical and computational tools for research across various disciplines. The program addresses the role of advanced computational techniques in mathematical modeling and simulation, covering applications in computer chip manufacturing, battery modeling, climate change, and more.
The mission of the program is to train Ph.D. students at UC Berkeley to use and manage scientific data, whether analyzing complex physical systems or employing statistics, machine learning, and data visualization for information extraction. The CDSE program is committed to multidisciplinary education, supporting scientists, engineers, and technical specialists across departments like Computer Science, Mathematics, Chemistry, Mechanical Engineering, Astronomy, Neuroscience, and Political Science. Upon completion, students receive a "PhD in X with a Designated Emphasis in Computational and Data Science and Engineering."
Program activities include identifying and encouraging relevant courses, promoting student involvement in CDSE research activities, supporting short "boot-camp" education programs, conducting summer schools, seminars, and tutorials. The program collaborates with Lawrence Berkeley National Laboratory (LBNL) and addresses the increasing importance of modeling, simulation, and data analysis in various scientific, engineering, financial, and social science fields.
Details...
The College of Computing, Data Science, and Society (CDSS) sponsors a Designated Emphasis in Computational and Data Science and Engineering (DE-CDSE), a program that is committed to the development of new curricula and expanded programs aimed at development and propagation of the use of numerical and computational tools to further research across multiple disciplines.
This interdisciplinary graduate minor recognizes the integral role of advanced computational techniques for mathematical modeling and simulation in a range of fields for the analysis of complex physical systems, such as computer chip manufacturing, battery modeling, turbine design, aircraft prototype testing, climate change and star formation, among others.
The dramatic increase in computational power for mathematical modeling and simulation has led to the fact that scientific computing now plays a significant role in the analysis of complex physical systems, such as computer chip manufacturing, battery modeling, turbine design, aircraft prototype testing, climate change and star formation, to name a few. More recently, too much data has become another compelling problem: radio telescopes, DNA sequencers, particle accelerators, sensor networks, social networks and the internet all collect much more data than humans can analyze and understand. In this case one needs to use statistics and machine learning, along with data visualization, to extract useful information from the data. The solutions to the mentioned problems share many mathematical, statistical and computational techniques in common.
Mission...
The program is open only to students who are already enrolled in a Ph.D. program at UC Berkeley.
The Designated Emphasis in Computational and Data Science and Engineering Program at the University of California, Berkeley trains students to use and manage scientific data, whether it is in analyzing complex physical systems or in using statistics and machine learning, along with data visualization, to extract useful information from the massive amount of data that can be collected from sensors today. The CDSE program is committed to the development of new curricula and expanded programs aimed at development and propagation of the use of numerical and computational tools to further research across multiple disciplines. To that end, the CDSE program will actively support the training and multidisciplinary education of scientists, engineers and technical specialists who are experts in relevant areas.
The CDSE program that crosses numerous disciplines, and participating departments include Computer Science, Mathematics, Chemistry, Mechanical Engineering, Astronomy, Neuroscience and Political Science, among many others. Upon graduation, the student receives a “PhD in X with a Designated Emphasis in Computational and Data Science and Engineering” on their transcript and diploma. This designation certifies that he or she has participated in, and successfully completed, a Designated Emphasis in addition to the departmental requirements for the PhD, and completion of the DE-CDSE will also be posted to the student’s transcript.
Activities include:
- Identifying existing and encouraging the development of new courses that best serve to educate computational science and engineering students
- Encouraging involvement in CDSE research activities by undergraduate and graduate students and postdoctoral researchers, and, potentially, by experienced professionals seeking to develop new skills that can benefit their careers
- Supporting formal short “boot-camp” education programs involving both live and web-based course offerings, with certificates if completed
- Supporting summer schools, seminar series, and tutorials
- Integration of these activities with Lawrence Berkeley National Laboratory (LBNL)
A great many fields of science, engineering, finance and social science are embracing modeling, simulation, and data analysis as necessary tools to advance their fields. Sometimes this is driven by the need to perform simulations of systems that cannot easily be directly measured, and sometimes it is driven by the increasing generation of large data sets that require extensive computation to understand. Both need to take advantage of the computational power which comes from continuing advances in computer components and architectures, including parallel computing.
Resources https://data.berkeley.edu/decdse_resources...
By Law:...
ARTICLE 1: Purpose...
The Graduate Group in Computational Science and Engineering shall establish and administer a graduate program of instruction and research leading to the Designated Emphasis in Computational Science and Engineering. The Designated Emphasis is applicable to existing Ph.D. programs at the University of California at Berkeley.
The interdisciplinary and diverse nature of academic and research interests of the faculty participating in the Designated Emphasis (DE) provide the student with a broad scope of educational opportunities. The appointment of an Executive Committee to monitor the activities of the Computational Science and Engineering Graduate Group, as described in Article 3, assures rigorous training for all students pursuing the DE.
ARTICLE 2: Membership...
Membership in the Graduate Group is open to faculty of affiliated Ph.D. programs who are actively engaged in teaching and research in appropriate sub-fields and who wish to join the Graduate Group. New members are considered for membership in the Graduate Group upon submission of a curriculum vitae to the Chair, listing their qualifications and stating reasons for seeking membership, and subject to approval by the Graduate Group’s Executive Committee. Proposed members are nominated to the Dean of the Graduate Division for appointment by the Executive Committee. Membership shall be reviewed periodically by the Executive Committee, at which time inactive members may be recommended to the Graduate Dean for removal from the membership list.
ARTICLE 3: Executive Committee...
Election. The administrative leadership of the Graduate Group shall be vested in an Executive Committee consisting of five elected members, all of whom are members of the Academic Senate. Members of the Executive Committee will be elected from among the Graduate Group membership, following nominations made by the Nominating Committee.
To ensure interdisciplinarity, the committee membership shall consist of 1) one representative from the Division of Computer Science, 2) one representative from the Department of Mathematics, and 3) three representatives from other departments. Any faculty may nominate other faculty to these 3 slates of candidates (CS, Math, and other). All faculty get 5 votes (1 for the CS slate, 1 for the Math slate, and 3 for the other slate); the three faculty from the other slate getting the most votes shall be elected.
Nominations by any member of the Group will also be included in the slate of candidates submitted to the membership. Election of the Executive Committee is by mail ballot and shall be completed by at least 3 weeks before the end of the spring semester. Votes cast electronically via e-mail are acceptable. The term of office will be two years following selection. Election results shall be announced to the membership by mail. Newly elected members shall assume their duties on July 1. In the initial election following approval of the Group, the three members receiving the largest number of votes shall serve a two-year term, the others a one-year term. The existing Executive Committee shall fill interim vacancies in its membership by appointment.
Executive Committee Chair. The Executive Committee will choose and recommend one of its members to the Graduate Dean for appointment as Chair of the group. The Chair of the Executive Committee is the de facto Chair of the Graduate Group for the DE.
Responsibilities. The Executive Committee shall meet at least once every semester to discuss issues of concern to the Graduate Group. The principal responsibilities of the Executive Committee are:
- To review faculty membership for the Graduate Group and maintain a list of active members to be annually reported to the Dean of the Graduate Division. This includes the nomination to the Graduate Dean of a tenured faculty member as Head Graduate Adviser for the DE.
- To review affiliation of affiliated Ph.D. programs and make appropriate recommendations to the Dean of the Graduate Division.
- To appoint the standing committees described in Article 5 below and other ad hoc committees as needed.
- To conduct administrative and clerical matters related to the activities of the Graduate Group.
The Executive Director will be invited to attend Executive Committee Meetings as a non-voting member.
The Chair of the Graduate Group (and Executive Committee) shall represent the Graduate Group membership in official matters pertaining to the DE. The Chair of the Graduate Group and the Chairs of the affiliated Ph.D. programs shall meet at least annually to discuss administrative, instructional, and research resource needs. The CDSE Executive Director will attend the meetings to provide high-level executive advice and to help to secure funding and promote the CDSE program.
ARTICLE 4: Program Standards...
Admissions
Admissions criteria for the DE will be devised by the Admissions Committee and approved by the Executive Committee as described in Article 5. Criteria will be clearly articulated to applicants (e.g., through the DE’s website).
Curriculum
The curriculum of the DE will consist of specified courses that may be independent from, or an integral part of, the doctoral programs with which it is associated. These will be determined by the Curriculum Committee as described in Article 5. The curricular requirements of the Ph.D. program and of the DE must be met before the student takes the qualifying examination. Any changes in curriculum requirements must be reported to the Graduate Division for review and approval.
Qualifying Examination
All students must be admitted to the DE program before the qualifying examination. The qualifying examination must include an examination of knowledge within the DE. At least one faculty member of the DE must participate in the qualifying examination, as determined by the DE Graduate Advising Committee or its designate, as described in Article 5, and in consultation with the Head Graduate Adviser in the affiliated Ph.D. program. Satisfactory performance on the qualifying examination for the Ph.D. will be judged according to the established rules in the affiliated program.
Dissertation
The dissertation topic shall incorporate study within the DE. The dissertation committee must include at least one faculty member of the DE, as recommended to the Head Graduate Adviser of the student’s home department or program by the DE Graduate Advising Committee or its designate, as described in Article 5, to ensure that the dissertation contributes in a significant way to the interdisciplinary field of the DE.
Degree Conferral
The DE will be acknowledged solely in conjunction with the Ph.D. in an established Ph.D. program and will be signified by the transcript and diploma designation “Ph.D. in Name of Student Major with a Designated Emphasis in Computational Science and Engineering.”
Program Evaluation
Evaluation of the academic quality of the DE will be conducted within the course of Graduate Council program review.
ARTICLE 5: Other Committees...
The following four committees will be appointed by the Executive Committee and have the following responsibilities:
Admissions Committee
- To establish student admissions criteria, to be approved by the Executive Committee, and review them for appropriate modifications.
- To review, and recommend for approval to the Head Graduate Adviser, graduate student applications to the DE.
- To collaborate with the affiliated Ph.D. programs in the recruitment of graduate students.
Curriculum Committee
- To review course requirements for the DE and make appropriate additions and deletions to the list of relevant courses. Changes in curriculum requirements must be reported to the Graduate Division for review and approval before implementation.
- To review Qualifying Examination and Dissertation requirements and make appropriate modifications. Requirement changes must be reported to the Graduate Division for review and approval before implementation.
Nominating Committee
- To propose a slate of candidates for open positions on the Executive Committee in the annual election.
Graduate Advising Committee
- To establish and review the DE’s mechanisms for advising students in adherence to standards of scholarship for the DE from pre-advancement until completion of DE requirements in the program, and to assess the quality of the program’s student advising.
- To assist the Head Graduate Adviser in the following duties of that position:
- To review and approve individual student course selections and make appropriate recommendations before approval.
- To review student Qualifying Examination and Dissertation Committees to ensure that the DE is represented by a core faculty member for nomination to the Dean of the Graduate Division for approval.
- To review and approve any variations from the DE standard Curriculum in individual student cases.
ARTICLE 6: Meetings...
At least one meeting of the general membership of the Graduate Group shall be held each year. The Chair shall call a special meeting whenever requested by written notice from five or more members. Meetings shall be conducted in accordance with generally accepted procedures. At meetings, twenty-five percent (25%) current members of the Graduate Group shall constitute a quorum. Minutes of each meeting shall be the responsibility of the Chair and shall be distributed to the membership and to the Dean of the Graduate Division following the meeting.
ARTICLE 7: Requirements for Ph.D. Programs to be DE Affiliates...
Existing Ph.D. programs of the University of California, Berkeley, may elect to offer the Designated Emphasis in Computational Science and Engineering as a program option by submitting a letter of intent to the Chair of the Graduate Group in Computational Science and Engineering DE, signed by the chair of the respective Ph.D. program. Affiliated Ph.D. programs must demonstrate commitment to training and research in the area of the DE, including the following commitments:
- Affiliated Ph.D. programs shall give the DE prominence at least comparable to that of a single faculty member in the program in brochures and website advertisements.
- Affiliated Ph.D. programs agree to either distribute DE materials (e.g., the program brochure) with all corresponding affiliated Ph.D. program materials, or provide such materials upon request.
- Applications of students indicating interest in the DE shall be marked for review by members of the DE Graduate Group in the respective Ph.D. Programs.
- Applicants indicating interest in the DE who are invited by the affiliated Ph.D. program to visit the campus (e.g., for interviews) will be encouraged to meet with members of the DE, including those outside of the affiliated Ph.D. program.
The Chairs of the affiliated Ph.D. programs agree to meet at least annually with the Chair of the DE Graduate Group to discuss administrative, instructional, and research resource needs. The CDSE Executive Director agrees to attend the meetings to provide high-level executive advice and to help to secure funding and promote the CSE program.
ARTICLE 8: Amendments...
Approval of changes in these by-laws shall require a two-thirds majority of the votes cast. Proposed changes shall be submitted to the membership of the Graduate Group by electronic mail ballot or for vote at a meeting, provided that written notice of the proposed changes is submitted to the members at least one week prior to the date of the meeting. Votes cast electronically via e-mail are acceptable.
Student registration...
The core thesis of the DE-CDSE program is that students need training in numerical analysis, scientific computation, and an application area. The DE-CDSE approach is as follows:
- It operates under the assumption that there is a large, diverse, set of students who are interested in CDSE, but that it may not be their primary core area of study.
- It targets students who may be interested in large-scale data analysis, not just simulation.
The DE-CDSE is designed for all these students and is open only to students who are already enrolled in a Ph.D. program at UC Berkeley. Essentially, Berkeley’s DE-CDSE program is meant to build on the student’s graduate expertise in a particular field and augment it with a focused addition on computational science.
Student Checklist: ...
Requirements for Ph.D. Students
Students must be approved to add the DE-CDSE before taking their qualifying exam. Submit your application at least 1 month in advance of your anticipated qualifying exam date in order to ensure enough time for processing and acquiring all relevant approvals.
Students must have training in the following three CDSE components from the approved course list (all 3 courses must be taken for a letter grade):
- 1 graduate course in applied math/numerical analysis or statistics/data analysis (Category 1)
- 1 graduate course in high-performance computing (Category 2)
- 1 graduate course in an application area that utilizes the above tools in a significant way (Category 3)
- Submit a complete online application, including:
- Curriculum Vitae (CV)
- Transcripts (most recent copy of all undergraduate and graduate transcripts)
- Letter of Recommendation from your advisor
- If your advisor is not a participating DE-CDSE faculty member, please include an additional letter of recommendation from a sponsoring DE-CDSE faculty member
- One-page statement about why you are applying to the program
Students must have a DE-CDSE component in their qualifying exam, with a DE-CDSE faculty member on the exam committee.
At least one member of the DE-CDSE faculty must be on the dissertation committee.
NC State Web...
MASTER OF SCIENCE IN DATA SCIENCE AND ENGINEERING...
Mission...
The MS in Data Science and Engineering is an interdisciplinary graduate program designed for students who seek to acquire computational and data science and engineering skills that allow them to tackle difficult problems in big-data, big-computation, and big systems.
The MS program offers its students various options:
- A Course-Only option designed to help students who want to graduate in one year.
- A Systems-Engineering concentration under the Course-Only Option designed for students and professionals who would like to hone systems-level thinking in addition to acquiring important computational skills
- A Project Option designed to give students a wide range of computational and data science and engineering skills and to enable them to graduate in 1 or 1.5 years.
- A Thesis Option designed to give students computational skills as well as research skills on tackling open-ended problems.
All these options will be an excellent preparation for our PhD in Computational Data Science and Engineering program.
Courses...
MS in Data Science and Engineering - Core Courses
- CSE 620 Introduction to Computational Software Tools
- CSE 704 Data Processing and Visualization
- CSE 708 Data Analytics and Engineering Applications
- CSE 817 Fundamentals of Big Data Analysis
PhD in Computational Data Science and Engineering - Core Courses
- CSE 702 Computational Methods for Algebraic Systems
- CSE 703 Programming for Scalable Computing Systems
- CSE 801 Computational Statistics
- CSE 804 Computational Modeling and Visualization
Other graded CSE Courses
- CSE 620 Introduction to Computational Software Tools
- CSE 701 Applied Probability and Statistics
- CSE 720 Research Computing Environments
- CSE 750 Concepts in Computational Data Science and Engineering
- CSE 803 High-Performance and Scalable Computing
- CSE 802 Computational Methods for Differential Equations
- CSE 885 Special Topics
- CSE 805 Machine Learning and Data Mining
- CSE 806 Computational System Theory
- CSE 815 Bioinformatics
- CSE 816 Concepts in Multi-Scale and Multi-Physics Modeling
- CSE 817 Fundamentals of Big Data Analysis
- CSE 826 Modeling and Simulation of Physical Systems
Other CSE Courses
- CSE 792 Graduate Seminar
- CSE 793 Master’s Supervised Teaching
- CSE 794 Master’s Supervised Research
- CSE 796 Master's Project
- CSE 797 Master’s Thesis
- CSE 799 Continuation of Master’s Thesis
- CSE 993 Doctoral Supervised Teaching
- CSE 994 Doctoral Supervised Research
- CSE 997 Doctoral Dissertation
- CSE 999 Continuation of Dissertation
Penn State Web...
Mission...
The mission of the Institute for Computational and Data Sciences is to build capacity to solve problems of scientific and societal importance through interdisciplinary, cyber-enabled research. As computation and data science become increasingly vital modes of inquiry, we enable researchers to develop innovative computational methods and to apply those methods to research challenges. Specifically, we:
- foster a collaborative, interdisciplinary scholarly community focused on the development and application of innovative computational methods;
- expand participation in interdisciplinary research through strategic investments and effective outreach; and
- provide a vibrant world-class cyberinfrastructure by maintaining and continually improving hardware and software solutions and technical expertise.
Vision...
ICDS will expand its role as an international leader in advancing cyberinfrastructure along with computational and data-driven methods and in driving their application to interdisciplinary research. We will use our expertise coupled with our state-of-the-science research infrastructure to support cyber-enabled interdisciplinary collaborations and attract the world’s best researchers. These researchers will form a vibrant intellectual community empowered to use the latest and most effective computational methods to make transformative discoveries for science and society.
MIT Web...
Summary...
The interdisciplinary doctoral program in Computational Science and Engineering (PhD in CSE + Engineering or Science) at MIT offers students the opportunity to specialize at the doctoral level in a computation-related field within eight participating host departments. These departments include Aeronautics and Astronautics, Chemical Engineering, Civil and Environmental Engineering, Earth, Atmospheric and Planetary Sciences, Materials Science and Engineering, Mathematics, Mechanical Engineering, and Nuclear Science and Engineering.
The program emphasizes thesis research activities focused on developing new computational methods or applying advanced computational techniques to significant problems in engineering and science. The research is expected to have a strong disciplinary component relevant to the host department. The program is jointly administered by CCSE and the host departments, requiring applicants to gain approval from both committees for admission. Admitted candidates must fulfill the host department's degree requirements, with specific deviations related to coursework, thesis committee composition, and thesis submission outlined on the CSE website. The CSE degree involves a course of study comprising five graduate subjects in Computational Science and Engineering. Applicants can find more details about the application process, requirements, and deadlines on the CSE website
Other details...
The Center for Computational Science and Engineering (CCSE) offers a master's degree and two doctoral programs in computational science and engineering (CSE)—one leading to a standalone PhD degree in CSE offered entirely by CCSE and the other to an interdisciplinary PhD degree offered jointly with participating departments in the School of Engineering and the School of Science.
While both programs enable students to specialize at the doctoral level in a computation-related field via focused coursework and a thesis, they differ in essential ways. The standalone CSE PhD program is intended for students who intend to pursue research in cross-cutting methodological aspects of computational science. The resulting doctoral degree is awarded by the Schwarzman College of Computing. In contrast, the interdisciplinary CSE PhD program is intended for students who are interested in computation in the context of a specific engineering or science discipline. For this reason, this degree is offered jointly with participating departments across the Institute; the interdisciplinary degree is awarded in a specially crafted thesis field that recognizes the student’s specialization in computation.
At the time of application, students are expected to declare which of the two programs they are interested in. Admissions decisions will take into account these declared interests, along with each applicant’s academic background, preparation, and fit to the program they have selected.
Applicants interested in an advanced degree in computer science should instead apply for admission to the Department of Electrical Engineering and Computer Science.
Further details: MITsub
New Jersey Ins Tech (2nd case) Web...
The CDS&E-MSS program accepts proposals that engage with the mathematical and statistical challenges presented by (1) the ever-expanding role of computational experimentation, modeling, and simulation on the one hand, and (2) the explosion in production and analysis of digital data from experimental and observational sources on the other. The goal of the program is to promote the creation and development of the next generation of mathematical and statistical software tools, and the theory underpinning those tools, that will be essential for addressing these challenges.