| Number 11
||University of Minnesota
Minnesota Center for Industrial Mathematics
INDUSTRIAL PROBLEMS SEMINAR
The MCIM/IMA Industrial Problems Seminar for the academic year 2003-4 featured 16 speakers representing various industries. Topics for the seminars were quite varied as well. Richard Derrig spoke about a mathematical model for insurance fraud detection. Apo Sezginer from Invarium, a Silicon Valley startup, spoke about how to advance lithographic methods in chip making. Former MCIM student Scott Shald gave a presentation about Coherent Technology’s effort in detecting leaks in gas pipelines. Many of the talks are recorded and available in streaming media from the IMA web site.
Seven students went on internships during the summer of 2004. Tara Rangarajan worked with medical researchers at the Hennepin County Medical Center on a method for measuring brain bloodflow in trauma patients. Thomas Hoft’s project at Coherent Technology was on scattering of light by aerosols. Chuan Xue developed a signal processing algorithm for data smoothing and interpolation for Medtronic. Other projects are described in MCIM’s web page.
STRENGTHENED COLLABORATION WITH MEDTRONIC
Cooperation between MCIM and leading medical device maker Medtronic, based in Fridley, MN, dates back to the summer of 1996 when the company started engaging mathematics graduate students as interns. Over the years, students have worked on projects as diverse as analysis of data for health management, to computing stresses in wire-rope cables. During the summer of 2004, Professor Carme Calderer, together with graduate students Brandon Chabeau and Hang Zhang, started a collaboration with Medtronic scientists on a materials modeling project. Both students are supported as research assistants in the School of Mathematics by a research grant from Medtronic. The project is likely to lead to scientific publications, proprietary intellectual properties, as well as two PhD theses for the students.
Medtronic expressed its appreciation for the value MCIM brings to the company by making a donation of $50,000 in April 2004.
|The accompanying photograph, taken at the Campus Club, shows Drs. Becky Bergman, VP for Research, and Darrel Unterecker, Director for Research, at Medtronic, with Larry Gray, Carme Calderer, Fernando Reitich, and Fadil Santosa.
Gilad Lerman, Assistant Professor in the School of Mathematics since Fall 2004, is a new faculty associated with MCIM. He received his PhD from Yale University and was postdoc at the Courant Institute for Mathematical Sciences at New York University before joining the department. Gilad works in several areas including harmonic analysis, multiscale data analysis, and learning theory. He is involved in many of the center’s activities and interacts with industry visitors as well as students.
INTERVIEW WITH YI-JU CHAO (PHD 1999)
|Yi-Ju Chao, who left after her PhD in 1999, moved back to Minnesota about a year ago. During her studies at the University of Minnesota she took part in the activities offered by MCIM. She was an intern at Medtronic in the summer of 1997, and finished her MS degree in Mathematics with emphasis in Industrial and Applied Mathematics in 1998. In the summer of 1998 she did an internship at Motorola. Both her MS and PhD thesis topics were based on the work she did as an intern. Currently a partner in the firm Morton Consulting Corporation, she recently spoke with Fadil Santosa about her experience after leaving the School of Mathematics.
Fadil Santosa: You were a student on the regular Mathematics track when you started here in 1994. What made you decide to take an internship?
Yi-Ju Chao: I heard a lot of buzz from other students about the industrial internships that were offered by MCIM. A lot of the students that have done it really enjoyed the experience. I also felt that it was the easiest way for me to find out what it would be like to work in industry.
FS: What was your first internship?
YC: I was an intern at Medtronic. I worked with Teresa Ruesgen on developing a predictive health model from heart-failure patient data.
FS: Did you feel you were well prepared for the project?
YC: Not at all. In the beginning, I was just trying to use existing mathematical tools to solve the problem. I found that I needed to understand the problem first. So, I read a lot of medical literature, and then developed a model based on the data.
FS: Do you think that what you did there was valuable to Medtronic?
YC: I remember the final presentation I made at Medtronic. The people in the audience asked a lot of questions, and were very excited. They expressed interest in continuing the research I started there.
FS: You must be happy about your contribution.
YC: Well, overall I felt I was ignorant about medical science and wished to learn more. But on the other hand, I did feel some satisfaction. It’s not the same satisfaction that I feel after proving a theorem. It’s different. I felt that I have used mathematics to answer a relevant problem from the real world. I ended up writing a Master’s thesis based on this work.
FS: Your second internship was at Motorola, and the research you did there ended up as part of your PhD thesis. How did you get connected with Motorola?
YC: I met Dr. Philip Fleming when he came to speak at the IMA Industrial Problems Seminar. You thought that there was potentially a good match and introduced me to him. I had an extended discussion with him about his research afterwards. Later he invited me to work with his team at Motorola as a summer intern.
FS: What was the project?
YC: Actually, it was quite a mathematical project. Phil Fleming, who was trained in mathematics at the University of Michigan, had posed a problem in queueing theory arising from communication networks. I was able to identify it as a mathematical problem. What I had to do was to prove that a certain type of semi-martingale weakly converges to a diffusion process. Much to Phil’s surprise, I solved this problem, which back then, was an open problem, within a few weeks.
FS: Did you publish this work?
YC: Yes, it has appeared in Queueing Systems, and it constitutes a major part of my PhD thesis, which was supervised by Professor Nicolai Krylov.
FS: How do you think you ended up with a job offer from Motorola after you finished your PhD?
YC: I always believe people are hired for a combination of reasons including also non-technical aspects. I think my boss believed that I could be trained and made valuable to the company even though I knew little about engineering. I also showed enough willingness to learn the engineering system that they were working on at the time. The project I described wasn’t the only thing I did at Motorola during my internship. After I finished it, I did some analysis of network data with heavy tailed distribution. The second project allowed me to work with other people on a problem which was very different, more practical. It proved to the people there that I could do more than solve math problems.
FS: What was your job title at Motorola?
YC: Lead software engineer. I guess that was the closest title they had for applied mathematician.
FS: How long did you work there and what did you do?
YC: I worked there for a little more than 18 months, and then moved to New York with my husband. When I first got to Motorola, I was given a tour of the projects they were doing. My first project was related to analysis of heavy tailed distribution, which was a hot topic during the Internet boom. But I soon realized that what I was doing had little to do with real product development. So I took on a project involving a new wireless network standard, doing performance analysis of wireless communication systems. This project later evolved into network architecture design. This turned out to be good training as it led to my second job.
FS: And where was that?
YC: The company is called InterDigital Communications Corporation. It is involved in wireless network design and wireless technology platforms. I was there for more than two years, developing network architecture designs, and representing the company in a standard organization of next generation mobile wireless networks.
FS: Now you are back home in Minnesota, and you established a consulting firm. What’s it like to have your own business?
YC: Having your own business is both exciting and challenging. We started out by doing consulting in financial engineering. We are now more focused on doing consulting in distributed control. Working for yourself you have a lot more freedom to choose, but it can be a challenge because your choice must make good business sense.
FS: It’s been over 5 years since you finished. In that period you have gone through several industries. What lessons have you learned that you could share with our students?
YC: Be prepared to keep learning when you leave school and enter the workforce, but not the same way as you have done in school. Mathematical skill is only a small part of what you need to succeed in industry. In the work place, your mathematical training will give you some advantages over your colleagues: one is the ability to reason rigorously and another is the ability to think deeply. But you will be at a disadvantage in other ways: one is ignorance of technology and the reality of the business world, two is lack of training working as part of a team, and three is lack of experience communicating with non-mathematicians. While your mathematical ability may get you a job in the first place, you have to improve yourself in the areas where you are weak. Someone who works hard to learn these other skills and who is a fast learner can advance very quickly.
FS: Should we do more training in non-mathematical skills such as those you described?
YC: First of all, the good news is that university does train students in mathematics very well and industry cannot provide this type of training that leads to rigorous and deep abstract thinking. These skills really do give an edge to mathematicians in industry. On the other hand, universities are not better than industry when it comes to teaching these other skills. The best approach is for the university to concentrate on doing what they do best in the classroom while providing students opportunities to interact with industry to develop other skills. I think I benefited a lot by talking to industrial speakers who visit the department, and of course, from the internships that I did. I recommend internship to every student, whether they will end up in academia or in industry.