Machine Learning Pathway
What do merging and important hardware platforms such as autonomous vehicles, robotics and the Internet of Things (IoT) all have in common? They all depend on machine learning, a branch of artificial intelligence and computer science that focuses on the use of data and algorithms to imitate the way humans learn. Machine learning is revolutionizing the ways in which hardware platforms are being used to collect data to provide new insights and advance fields ranging from medicine to agriculture to aerospace.
The Machine Learning pathway connects computing, machine learning and data science together by explaining how hardware and sensor design enables data collection at-scale; then how to store, process and clean the data; and finally how to train and test machine learning models using collected data to produce new insights and model predictions applicable to real-world problems. This is an excellent pathway for students who want to learn how to combine hardware and software into projects that address needs in a wide range of fields.
This pathway is a good fit for students who are interested in:
- Learning fundamental concepts of how to manage datasets, which includes pre-processing, cleaning, feature extraction, selection and iteration, and how to train and test models using cross-validation, supervised, unsupervised and reinforcement learning
- Learning how to distinguish between use cases for core techniques such as classification versus regression, and causation versus correlation
FAQs
Does a student need a graduate degree specializing in this area to be marketable to industry?
No. However, while a student with a bachelor’s degree can easily find employment, a master’s degree opens up a noticeably wider range of positions in this area. A doctoral degree is generally required for research or teaching.
How can a knowledge of machine learning be applied in the real world?
Developments in computing are leading to advances in hardware, which enable machine learning to be run both on the device (for fast, lightweight processing) and in the cloud (using big data and graphics processing units). Real-world examples include autonomous vehicle navigation, Alexa, Google Maps for routes with the least traffic, Roomba and Boston Dynamic robots.
Does machine learning touch on global impact, equity and/or quality of life?
Yes. Advances in computing, alongside machine learning, can be used to make an impact locally, nationally and globally. For example, machine learning can contribute toward helping to better predict and improve population health, detect plastics that are polluting our rivers and improve the sustainability of crops.
Areas of Impact
Air and Space
Machine learning is used to improve efficiency and accuracy for robots on-board the International Space Station, such as Astrobee, to collect data in microgravity conditions while interacting with astronauts. Drones deployed in search-and-rescue missions can use integrated sensor imaging along with machine learning to improve their search time and accuracy of locating people who are lost or injured.
Computing Data and Digital Technologies
Google, Amazon and Microsoft have a new focus on an emerging area of development in the field of Green AI, which explores optimizing hardware and software to minimize power consumption. The development teams aim to detect drift in machine learning models, thereby alerting users that they need to re-train their models, which will lead to more accurate predictions and consume less energy.
Environmental Sustainability and Energy
Machine learning is contributing to the development of next-generation weather balloons capable of helping to predict and monitor wildfires.
Health and Medicine
Machine learning is part of a wide range of health and medical applications such as developing devices to improve the lives of an aging population, repurposing smartphone sensors for mobile health applications and improving the accuracy of detecting faint lines in rapid diagnostic tools.
Infrastructure, Transportation, and Society
Machine learning is being used in the development of sensor arrays for use in Smart Cities, as well as improving the efficiency of flights, delivery of goods and services, and supply chain management.
Robotics and Manufacturing
Hardware development in conjunction with machine learning algorithms improves autonomous robot efficiency, effectiveness and usability, enabling better manipulation and navigation of robotics on land, sea and in the air.
Related Career Paths
Students graduating with a focus in machine learning will be qualified for jobs in hardware and software engineering at leading tech companies such as Google, Amazon, Microsoft, T-Mobile, Tesla, Meta, Nvidia and Apple, as well as startup companies.
On-the-job tasks for graduates with a focus in Machine Learning include:
- Training large machine learning models using optimization techniques, testing how they generalize information and finding ways to improve the models.
- Applying trained machine learning models for various applications and tasks, such as classification, detection, segmentation and recommendation.
- Applying rigorous mathematical modeling for applications, identifying and designing suitable optimization algorithms.
Machine Learning Courses
When planning for courses, review projected course offerings here and be sure to check all course prerequisites (course titles below link to the catalog course description, which includes prerequisite information).
These courses are suggested for those following the Machine Learning pathway but are not required to complete the BSECE degree program:
EE 342 — Signals, Systems, and Data II
In this course, students will review basic signal processing concepts. Topics include two-sided Laplace and z -transforms and connection to Fourier transforms; modulation, sampling and the fast Fourier transform; short-time Fourier transform; and multirate signal processing. Applications covered include inference and machine learning. Students will also participate in a computer laboratory.
EE 443 — Machine Learning for Signal Processing Applications
Learn about the application of machine learning and deep learning algorithms to real-world signal, image and video processing problems using cloud computing with central, graphics, and tensor processing units (CPU/GPU/TPU). Course topics include covering the characteristics of multi-dimensional signals and systems, unsupervised and supervised learning, deep learning convolutional neural networks, generative adversarial learning, open long-tailed recognition, and object detection and segmentation.
EE 445 — Fundamentals of Optimization and Machine Learning
Introduction to optimization and machine learning models motivated by their application in areas including statistics, decision-making and control, and communication and signal processing. Topics include convex sets and functions, convex optimization problems and properties, convex modeling, duality, linear and quadratic programming, with emphasis on usage in machine learning problems including regularized linear regression and classification.
EE 469 — Computer Architecture I
How does the machine code produced by a compiler translate into computation by a processor? How can we improve the performance of a processor, and what are the trade-offs that must be made? These questions and many more are answered by this course, as students receive an initial exposure to computer architecture and design their own processor in Verilog Hardware Design Language, an industry standard for hardware description.
CSE 373 — Data Structures and Algorithms
Fundamental algorithms and data structures for implementation. Techniques for solving problems by programming. Linked lists, stacks, queues, directed graphs. Trees: representations, traversals. Searching (hashing, binary search trees, multiway trees). Garbage collection, memory management. Internal and external sorting.
CSE 374 — Intermediate Programming Concepts and Tools
Covers key software development concepts and tools not in introductory courses. Concepts of lower-level programming (C/C++) and explicit memory management; techniques and tools for individual and group software development; design, implementation, and testing strategies.
CSE 415 — Introduction to Artificial Intelligence (*does not count toward Adv. ECE Electives)
Principles and programming techniques of artificial intelligence: LISP, symbol manipulation, knowledge representation, logical and probabilistic reasoning, learning, language understanding, vision, expert systems, and social issues.
Capstone
The Machine Learning pathway offers the potential to pursue projects that synthesize concepts in computing, machine learning and data science through the ENGINE Capstone series of courses (see description below). Additionally, the Embedded Systems Capstone (see below) can allow students to explore machine learning projects within an embedded systems framework. Examples of recent machine-learning-oriented ENGINE Capstone projects include: A Lockheed-Martin sponsored project to design, build, and test a scaled self-driving vehicle; a JPL sponsored project exploring autonomous distributed robotic exploration, and a drone automated payload and battery swap/charging project.
EE 475 — Embedded Systems Capstone
In this capstone class, students will work in teams to apply knowledge they gained in EE 474, Introduction to Embedded Systems, and from other previous ECE courses to prototype and build a substantial project that mixes hardware and embedded software and communication. Students often build projects in specific application areas that include but are not limited to health, robotics, Internet of Things (IoT), and smart systems. Students will also hear from experts in embedded systems to learn about emerging platforms, trends and job opportunities/prospects.
EE 497 (winter quarter) and EE 498 (spring quarter) — Engineering Entrepreneurial Capstone (ENGINE)
The Engineering Entrepreneurial Capstone program (ENGINE) is the culmination of a student’s electrical and computer engineering education at UW ECE. The program provides a unique opportunity for students to develop skills in collaborative systems engineering, project management, and most importantly, working in teams on real-world problems from industry-sponsored projects. The program is overseen by UW ECE faculty and students are guided by practicing engineers. The course culminates in a showcase of student projects, which is attended by industry sponsors and held at the end of spring quarter every year.
Crossing Paths
Students studying Machine Learning should also consider the following customizable pathways:
Enriching Your Path
The following courses are also recommended for those following the Machine Learning pathway: