Courses for NRT Trainees

Core courses: The core courses offered in the graduate curricula spread across various participating UD departments will serve as a strong disciplinary foundation after which the NRT trainees will be introduced to cross-disciplinary training through two elective courses.

NRT specific courses: Going beyond these core and interdisciplinary electives, to provide a truly convergent technical training in HPC and data-driven materials informatics, we realized the need to develop two new NRT courses.

1. NRT Hackathon Course: The hackathon-style course is mandatory for our NRT (funded and unfunded) trainees after the trainees complete their disciplinary core classes and the two cross-disciplinary electives, and before the trainees’ summer internship/teaching workshop. This hackathon course will be cross-listed as an elective in all the participating trainees’ departments and will be broadly available to graduate students outside of the NRT program. This course, motivated by the successes of GPU hackathons led by co-PI Chandrasekaran, will involve teams of 3-4 trainees from different disciplines working together to solve (over the semester) soft materials (MAT) problems submitted by academic labs, industry, and national laboratory using high-performance computing (HPC) and/or data science (DS) tools. Researchers in industry and national lab partners who will provide such MAT problems will also serve as mentors to the teams during this hackathon course.

This course will be co-instructed by two NRT core-team faculty with convergent expertise. The first offering of this course will be co-instructed by Jayaraman, Chandrasekaran, and Brockmeier.  During the NRT hackathon course, each MAT problem will be tackled by a team of 3-4 trainees/students with 1 from HPC background, 1 from DS and 1 from MAT. This team composition will force the students to learn to communicate with researchers outside of their primary expertise. The  co-instructors and industrial/national lab partners mentors will guide the teams through teamwork, collaboration, and oral & written communication exercises.

The trainees and enrolled students will a) develop and sharpen their technical skills in MAT, HPC and/or DS applied towards practical problem solving; b) learn to communicate across research expertise; and c) develop strong interpersonal negotiation, teamwork, and collaborative skills.

 

2. NRT Capstone Design Course: To sustain the interdisciplinary training beyond their two years in the NRT program, and to better infuse this training into their long-term convergent research, in the Fall semester of their 3rd year, each trainee will enroll in this ‘capstone design’ course co-instructed by two NRT team faculty (rotating annually).

The trainees will integrate their ongoing convergent research work along with the gained technical and professional skills into a forward-looking three-year (post-NRT) plan of work for their doctoral thesis. The trainees (and other enrolled students) will develop a detailed plan of work for their own convergent research thesis topic (goals, specific aims, approach, expected outcomes, preliminary results, anticipated challenges, and alternate plans). They will communicate various parts of this plan of work (oral and written) to fellow students during the semester and to a panel of academics and external collaborators, industry, and national laboratory mentors at the end of their semester.

During their time in this new Capstone Design course, trainees and other enrolled students will experience the benefits of teamwork, peer-mentoring, and have multiple opportunities to develop and practice their communication skills with a broad audience. The students in this course will work in pairs on their technical project and community outreach activity; the members of this pair will rotate twice during the semester to expose students to different technical backgrounds and thesis topics.

The students will be tasked to communicate with people outside of their class by creating one visually appealing image of their convergent research as such images can serve as “clear and engaging narrative to communicate complex scientific topics”. Students will engage with public and communicate to non-scientists by sharing these images on our NRT website, as well as by creating a research news bulletin (1/2 page of eye-catching content) to pin in coffee shop bulletin boards, and by having the whole class collaborate to create an annual newsletter highlighting the projects to share electronically with NRT trainees’ extended network; these will also be posted on the NRT website. We will have experts within UD department of communications give feedback to the trainees and students on their work. Through these communication exercises at various stages of this course, the trainees will practice how to explain in layman terms the potential benefits and broader impacts of their research to the stakeholders and the public.

 

Elective courses: Guided by their primary and secondary co-advisors, the trainees will select appropriate elective courses from the discipline complementary to their core expertise. For example, a trainee from MSE/CBE/CBC/BME department would select DS and/or HPC electives, while a trainee from CIS/ECE/DS will select MAT focused electives that also cover relevant HPC/DS tools for MAT. The NRT technical training co-directors will work with graduate program directors in various departments to get these electives approved in their graduate curriculum.

The following are existing NRT relevant UD electives that the NRT trainees could choose from during their traineeship:

1) ELEG 817 -‘Large scale machine learning’ (Brockmeier): This course prepares students for using ML and DS in a variety of domains by focusing on how to scale up approaches for analyzing real-world data sets. The topics involve computational and statistical aspects from theoretical and practical perspectives. The formative assessments provide exercise in problem formulation, critical reading of literature, algorithm implementation, abstract writing, experimental design, and peer review. MAT domain problems will be included in this course to align this elective with this NRT program.

2)  MSEG 667- ‘Organic electronics’ (Kayser): This course covers the fundamentals of organic electronics (polymers and small molecules), their synthesis, design principles, structure-property relationships, materials processing and device fabrication. The course includes a section on computational methods (atomistic and coarse-grained modeling and simulations) with guest lecturers which could be expanded further to align with the HPC+DS tools.

3) To be offered soon ‘Parallel programming and data science’ (Chandrasekaran): The success and impact of this undergraduate core course led Chandrasekaran and Stephen F. Siegel from CIS to put together a graduate level elective which will soon be offered. This course will cover the fundamentals of parallel concepts, patterns, methodologies, and programming models used for parallelization, parallelization of both multicore systems and systems with accelerators and will be expanded to integrate NRT training by studying programming challenges from MAT domain scientists.

4) ‘High-Performance Computing (HPC) 101’ (Eigenmann): This is an introductory course to HPC and DS that students across the disciplines take to learn basic principles, terminology, and technology needed to develop and run efficient computational and data science applications.

5) CHEG 867 ‘Molecular modeling and simulations of soft materials’ (Jayaraman): This course provides an overview of coarse-grained and atomistic models, molecular dynamics and Monte Carlo simulations in various ensembles, and applications of these models and simulations to study problems in chemical and materials sciences.

6) CHEG 867 ‘Process Systems Engineering: Mathematical Modeling and Optimization Principles’ (Ierapetritou): This course provides fundamental instruction on the mathematical programming techniques used in the solution of process design, synthesis and operations problems. This course will review techniques in optimization and mathematical modeling to study synthesis analysis, evaluation, and optimization of process alternatives, process operations involving planning and scheduling, uncertainty considerations on process design and operations.

All UD based data science related courses can be found here as well.

Graduate students can reach out to the indicated course instructors if they have additional questions about the above electives.