The Davidson College of Engineering is launching a new Masters of Science program in Artificial Intelligence (AI) in fall 2020. The first cohort of the MSAI program will be offered through Engineering Extended Studies with one of the College of Engineering’s off-campus corporate partners.

“With the launch of this program, our college continues to offer state-of-the-art degrees that drive today’s technological advances,” said Sheryl Ehrman, the Don Beall Dean of the College of Engineering. “Some people do programming and some build engineering systems: this program builds a bridge between the two groups, enabling traditional engineers to learn artificial intelligence technology and be able to apply it to their engineering professions.”

Xiao Su, chair of the Computer Engineering Department, said, “Although it is a relatively new field, artificial intelligence has already changed the world in many ways. There is a lot of potential in all areas of life and all fields, including finance and health care. The objective of our program is to fill in the gaps for people who would like to apply AI in their chosen profession, and know that they need preparation in order to succeed in the program.”

People who have math and programming backgrounds and experience, and ideally an engineering, science, mathematics, or computer science undergraduate degree, can use this Masters program to apply artificial intelligence technology to their engineering profession.

Students will primarily use Python programming in their courses and will work with TensorFlow and Pytorch–the two most popular libraries. They will learn regression, classification, decision trees, IU base algorithms, clustering, recommendation, neural network learning and deep reinforcement learning. “They will also learn GPU programming, which made deep learning possible,” said Su. “The combination of availability of GPU and the deep neural network is what ignites the current artificial intelligence evolution.”

“We thought of artificial intelligence in three layers,” Su continued. “The underlying layer is the algorithms to make the APIs and make the underlying technologies more efficient. The middle layer is “tool box” or API, and the top layer is the application layer. Graduate students in this program will be able to work with the middle and top layers. Students who select electives in machine learning, deep learning and reinforcement learning will be able to build tools and APIs that drive artificial intelligence. Other students who might be more application-oriented will be able to build engineering systems using AI technologies.”

Students will be expected to finish within two years. They will have immersive training in machine learning and artificial intelligence, as well as hands-on preparation and experience. They will learn the foundations, evolution and advancements in artificial intelligence, in order to understand not only the current technologies, but also to prepare to be part of the future of artificial intelligence.

“We pride ourselves on making this program hands-on,” added Dean Ehrman. ‘Hands-on’ is the thread that ties the whole program together, and that is what we are known for at Davidson College of Engineering.” Besides the final capstone project, each course has a project where students can work individually or in teams to build a system, prototype, or application.

“We plan to bring the program to the main campus starting Fall 2021,” said Su. “We want to explore our corporate partnerships first, which will allow us to get additional feedback and fine tune the enrollment process. We plan to gradually increase in size. There is a high demand for AI engineers!”

In a competitive district such as this illustration from the paper, the maximum support advantage of either party is at most 10%. These parameters can be selected by the redistributing commissions, including the number of competitive districts and the vote split.

The unfair and non-competitive political redistricting process known as gerrymandering could be solved using new mathematical models and computer implementation, according to a recent study by a San Jose State University research team that appeared in Computers & Industrial Engineering, an Elsevier publication.*

“Election and redistricting are political processes; we believe that software tools are needed to achieve political purposes,” said one of the paper’s authors, Dr. Jacob Tsao, a professor of Industrial Systems Engineering. “We present two mathematical optimization models to implement political fairness and competitiveness.” Both models are implemented with a case study of South Carolina.

There appears to be a growing movement against gerrymandering. In December 2019 a New York Times columnist observed that State courts in Pennsylvania and Virginia have thrown out gerrymandered maps, voters in Arizona, California and Michigan approved ballot measures to reduce gerrymandering, and similar measures could be on the ballot in 2020 in Arkansas and Oregon.

“Redistricting can be formulated and solved as a difficult purely mathematical problem, without any partisan biases; such efforts started as early as the 1960s but continued intermittently and sparsely,” wrote the San Jose State team in their study, hoping that their research will assist redistricting commissions and their staff members by developing a suite of mathematical models and their computer implementations to develop and optimize districting plans.

“Hopefully, this is a timely paper for the political redistricting work to occur across the United States in response to the ongoing 2020 US Census and in preparation for the 2022 election and beyond,” said Tsao. The team writing the paper included not only SJSU faculty and students but alumni as well. “The first author of this paper, Dr. Hongrui Liu, was an MS-Industrial & Systems Engineering (ISE) student about 15 years ago,” explained Tsao. “She and I published a journal paper in 2008 on the design of experiments, with an unconventional approach, based on her MS Project. She went on to earn a Ph.D. from the University of Washington in 2010. After working for several years in the energy industry, she joined SJSU ISE faculty in 2017.”

San Jose State’s math department is also thinking about politics, offering Math 10P (Mathematics in Politics) for the first time in spring 2020. The course description states: “Did you know that there’s a state in the U.S. where one party gets 52% of the vote, but nevertheless holds 75% of that state’s seats in Congress? Did you know that San Francisco, Oakland, Berkeley, San Leandro, and the state of Maine use a different system for counting votes than most of the rest of America?

“Believe it or not, the key to understanding all of that and more is math! In Math 10P, Mathematics in Politics, you’ll learn not only the mathematical secrets behind all of those phenomena, but also how, through politics, math impacts our daily lives and the future of America. Recommended for anyone with an interest in politics, current events, or in how math and politics affects our daily lives. Fulfills [the] Math/Quantitative Reasoning requirement.”

*“Mathematical models of political districting for more representative governments” by Hongrui Liu, Ayca Erdogan, Royce Lin (a SJSU MS-Industrial & Systems Engineering Student), and H.-S. Jacob Tsao. Computers & Industrial Engineering, an Elsevier publication. Volume 140, February 2020, 106265.