Education

Fall semesters: second-year graduate-level course in machine learning.

Spring 2023: second-year graduate course with industry proposed team projects focused on Computing and Data Science for Soft Materials Innovation (co-teaching with lead instructor Prof. Arthi Jayaraman)

Spring 2023: In development: a 300-level undergraduate course in geospatial data science (co-teaching with Prof. Jing Gao and Prof. Gregory Dobler

Spring 2022–: NRT-MIDAS NSF funded research traineeship, serving as Core Faculty and secondary advisor for trainees

Spring 2020, 2021, 2022: 300-level undergraduate course in signal processing and linear system analysis

Summer 2021: Independent study on adaptive filters

Spring 2021: Graduate Data Science Community Hour during Spring 2020

Graduate Course

Large-Scale Machine Learning ELEG/FSAN 817

Course Description: Introduction to the analysis and processing of massive high-dimensional data. Massive data sets generally involve growth not only in the number of individuals represented but also in the number of descriptive parameters of the individuals, leading to exponential growth in the number of hypothesis considered. Approaches to address these problems exploit concepts from statistics and machine learning.

Offered: Fall 2019, Fall 2020, Fall 2021, Fall 2022

The course serves as a capstone course in machine learning for graduate students, with the goal of readying students for machine learning and data science research. Computational and statistical scaling of machine learning is discussed from both theoretical and practical perspectives. The formative assessments provide exercise in key aspects of machine learning research: problem formulation, critical reading of literature, algorithm implementation, abstract writing, experimental design, and peer review. Summative assessment is based on a mid-term exam and final project.

2021 hyrid (in person/online synchronous) course enrollment: There were 9 students enrolled: ECE (2 Ph.D. | 4+1 B.S./M.S.), CS (5 Ph.D.), Bioinformatics Data Science (1 Ph.D.).

2020 online synchronous course enrollment: There were 17 students enrolled: ECE (9 Ph.D. / 2 M.S.), FSAN (3 Ph.D.), Computer Science (1 Ph.D.), and Physics and Astronomy (1 Ph.D.). One additional ECE Ph.D. student audited.

2019 course enrollment: There were 13 students enrolled: ECE (3 Ph.D. / 2 M.S. / 1 B.S./M.S.), Computer Science (5 Ph.D.), Economics (1 Ph.D.), Civil & Environmental Eng. (1 Ph.D., auditing).

Anonymous feedback from 2021 course evaluations:

• “Prof Brockmeier has done a great job of curating a higher-level data science course for grad students that helped inform me about the origins of the fundamentals of machine learning, the relevant aspects of probability and mathematics that relate to machine learning, and the state of the art methods that are popular right now. The class felt linear and started at a basic level where the instructor introduced fundamental concepts and basic mathematics, and slowly introduced papers that were important in the field. The weekly assigned readings were generally helpful and the homeworks were selected well and reflected material presented in the course.”
• “I very much enjoyed this class and learned a lot. I would describe the class as an introduction to reading ML research papers, which is an extremely useful skill. Professor Brockmeier selected some of the key research papers in statistical learning and would break down the background material and key concepts in class. For homework, we read these papers, much better equipped to understand the technical details. By the end of the semester, we were able to understand some of the key concepts in statistical learning and possess the skills to continue learning and researching for our course projects. This course also helped me in my research, as my advisor is not focused on ML. By being able to understand ML papers and what to look for when reviewing has opened doors for me, as I now feel like I would be able to publish in ML journals. Overall, it was a great course and I highly recommend it to any graduate student whose research touches upon ML”
• “I really enjoyed the class, […] our interaction with the instructor was having a major role in the educational process. The in-class atmosphere and flexibility (time and topics) are top-notch, everything is done to make the course fit as many ML-related majors as possible.”
• “I thought the course was very interesting, and definitely expanded my understanding of machine learning”
• “Clearly a subject matter expert. Communicated material well. I always felt like I could ask questions in class, office hours, or email. Professor showed that he was focused on actual learning beyond what is often demonstrated by other faculty.”
• “He included students in the class and was mindful of both those present in person and on Zoom. He made the class interactive and made sure everybody had a chance to speak their thoughts, questions and opinions about different models. He also fostered a good peer review process which allowed students to review our peer’s projects and provide constructive feedback. He laid out during class what constructive feedback looks like for this process”

Anonymous feedback from 2020 course evaluations:

• “The topics covered in the course are interesting and relevant. The instructor exhibits a deep understanding of the subject and has the capacity to motivate interest in the students.”
• “The instructor did a great job structuring the course and selecting the key concepts and methods among the vast amount of topics in large-scale machine learning. He is very knowledgeable and demonstrates an honest interest in our learning experience. The evaluation tools used (the midterm, the pop quizzes, the homeworks, and the final project) are well designed to accomplish our learning goals. I particularly liked that he made emphasis throughout the course in that the evaluation tools were more intended as learning tools, and offer us options to ease the natural pressure that the grades associated with those tools usually have.”
• “I like the discussions and that despite being online, the professor was capable of keeping me engaged even when the topics were not particularly easy to explain. I appreciate the use of a whiteboard instead of the slides, it gives flexibility to the class and it helps me to understand.”
• “He always wants us understand the topics. He supports lecture[…] interactions. He asks during lectures to make sure we are on the right track. He tried to link topics that are correlated and related to each other. He links the topics we are studying to a real world examples.”
• “I can see Prof. Brockmeier really wants students to learn, no matter whether a student is already a machine learning research[er] in a specific area or is still exploring the subject.”
• “Dr. Brockmeier is a great professor. He was well prepared for the classes and was capable of address almost every question that pops up during the course and always made a good connection between theory and Matlab examples.”
• “There will be timely comments and replies to every student’s questions and papers”
• “He was open to listening to the questions of all the students that wanted to ask and address those questions in a respectful way.”
• “He was always polite and gave everyone the opportunity to express their opinions and ideas regarding the topics that were explained.”
• “I rank him in the top 1 percentile for research ability, approachability and overall being a good person. I would take his class again and recommend it people.”

Anonymous feedback from 2019 course evaluations:

• “The course really dives deep into the mathematics and theory of the machine learning techniques. The organization of the course was nice.”
• “Course material was well ordered, well motivated, and useful.”
• “The breadth of topics studied in this class provided me with a better understanding of machine learning in general.”
• “Professor Brockmeier […] was able to connect abstract concepts from the class to multiple applications to motivate understanding”
• “This was not only the best course I took as a graduate student, but moreover, the most useful in the pursuit of becoming a machine learning expert.”
• “He was enthusiastic about delivering the contents of the course.”

Undergraduate Courses

Signals and Systems ELEG 305

This course examines continuous and discrete-time signals and systems at the introductory level.

Course description: Introduction to signals and systems, with an emphasis on time and frequency characterization of linear, time-invariant systems. Covers discrete and continuous time systems, sampling, and Fourier, Laplace, and Z transforms.

Offered: Spring 2020, Spring 2021, Spring 2022

Core requirement in the following programs:

• Electrical Engineering
• Computer Engineering and computer engineering majors.
• GIScience and Environmental Data Analytics

Technical elective:

• Chemical Engineering, Materials Science and Engineering, Mechanical Engineering

Prerequisite for subsequent courses:

• ELEG 306 - Digital Signal Processing
• ELEG 403 - Communication Systems Engineering
• ELEG 404 - Digital Imaging and Photography
• ELEG 418 - Digital Control Systems
• ELEG 492 - Radar Systems and Technology
• BMEG 441 - Biomechatronics†
• BMEG 330 - Biomedical Instrumentation†
  † As an alternative for BMEG 230

Honors Discussion Section: Students registered in Honors for this class will be involved in an additional discussion section, which introduces content beyond the general class: specifically, non-linear, adaptive, and random systems. They will also be required to complete a semester long project that reflects an application of such systems.

Unsolicited Feedback from 2021 Honors section:

• “I really enjoyed looking more into something I’m very interested in. It was my favorite Zoom meeting between all of my classes this semester. Connecting with what I focused on, […] Doing Honors research got me to realize that dream.”

Unsolicited Feedback from 2020 section:

• “I want to applaud your tenacity and optimism throughout the term. Despite this class being the most challenging course I’ve ever taken, I have enjoyed being your student.

Independent Study ELEG 466

“Adaptive Filtering as an Introduction to Machine Learning”

Offered: Summer 2021

Course description: Introduction to adaptive filter theory and application. Students will create a technical report that details the theory, implementation, and computational experiments with the application of adaptive filtering. Topics to be covered include review of delay lines, z-transform, multivariate signal processing, Gaussian white noise, correlation matrices, whitening transforms, least squares, convex optimization, Wiener filter, stochastic gradient descent, least mean squares (LMS), NLMS, common spatial patterns, and various types of recurrent memory. A comprehensive technical report with equations and figures will be typeset in Latex. The report will detail the problem formulations, equations, and vector graphics of block diagrams, plots of results. Software implementations will be made in MATLAB/Octave, Julia, python/Numpy, or python/Pytorch. Synthetic experiments will be conducted for noise cancellation, channel estimation, and machine learning.

Learning outcomes: While students typically learn theory and fundamentals of signal processing and filtering in ELEG305 “Signals and Systems” and probability theory and statistics related to signals in ELEG310, due to time constraints and curriculum design, students rarely have the opportunity to learn and apply adaptive filter theory that combines filtering, linear algebra, optimization, and statistics together. This combination is the foundation of machine learning theory and algorithms. This independent study will attempt to fill the gap in the current curriculum areas in theory and practice. A combination of synthetic data and real-world projects will be explored by students to master these concepts. Students will practice their understanding by describing the formulations and results (equation and paper typesetting, block diagrams, data visualization and presentation) in their own words as well as implementing and testing software for adaptive filtering and machine learning. In summary the students will learn the foundations and techniques of machine learning: linear and non-linear filtering, supervised learning, least squares, optimization, stochastic gradient descent, batch normalization, and experimental design (training, validation, and testing).

Experience: The course will solidify knowledge learned in other courses and provide the student with a diverse and broad understanding of adaptive filtering, and an introduction to machine learning. The course project will require tools learned in other academic classes (signal and systems, linear algebra, calculus, probability, and statistics) and will help the student in future classes (communication systems, machine learning, control theory) as well as research.

Expectations: Students will be required to complete a technical report which will be a culmination of most of the topics covered in class. Weekly revised versions of the technical report must be submitted to document learning progress until the end of the course. Each of these weekly submissions, will require an executive summary that overviews the specific learning activities, progress in understanding and experiments, obstacles, and interesting insights or useful skills that have been learned. Students can work in groups and collaborate in a civil and equal manner.