Meet the Faculty - Ali Anwar
Tell us about your journey to the University of Minnesota.
After getting my Ph.D. from Virginia Tech, I joined IBM Research Almaden as a research staff member. I did three internships during my Ph.D. program with IBM and they offered me a role during my last internship. I thought that it would be a good place for me to figure out what I really want to do after my Ph.D. IBM Research is a place where you get to experience both industry and academia. I got a chance to work with a lot of interns and students, and I really liked the mentoring part of my job and publishing papers while working with students. So that’s when I decided to pursue teaching full time.
I chose the University of Minnesota because they are very bright and active researchers. The faculty is really awesome at the University and I think it is a good fit for me. I work at the intersection of distributed systems and machine learning (ML), and it is hard to find a place where you can get good support for both of these areas. At Minnesota there is a really good systems group as well as great AI (artificial intelligence) and ML folks, so it is a great match for me.
Plus, during my interview process I learned about the Data Science Initiative and I think that is going to be a game changer for the University of Minnesota.
We would love to hear more about your research!
I work on developing systems that make modern applications run more efficiently. Modern applications can be machine learning, and cloud storage like Google Drive, Facebook, etc. All of that information is obviously not stored on individual cell phones; it is housed elsewhere but is still accessible to each individual. So I work on systems that make that process possible.
Recently I have been working more on machine learning, developing distributed systems that will allow you to train models with AI. My current project is federated learning, a subtype of machine learning. It’s a relatively new field and I am actively working on privacy-preserving federated learning on personal devices. Because of regulations such as GDPR, HIPAA and other liabilities, companies are electing to not take private data from a single device to somewhere central to train a model. They now prefer to train the model at the data source and take that model instead of the data. That way they can still get insights without violating privacy.
There is another level of federated learning called enterprise-level federated learning, which I focused on while I was at IBM. In bigger businesses like hospitals and banks, they train their own models with their own data because they don’t want to share their information with competitors. Then I work to combine these models into a generic global model that can be widely used without specific dataset for an individual company.
Currently I am working on federated learning in a resource-constrained environment where the computational capacity is very low. When you are working with data from a cell phone, there are battery, data and computational constraints that are different from working on advanced computers that are built for that sort of work. So I am focusing on tackling the differences in functionality between different devices.
What do you hope to accomplish with this work? What is the real-world impact for the average person?
Federated learning is already helping a lot in the medical field. Because of HIPAA, doing research on patient data is very difficult. Federated learning allows medical research centers to work with hospitals without the need to share data. It allows for more collaboration and helps these researchers make advancements in diagnosing diseases. It can also be used to put together a personalized medical plan for an individual based on millions of other users’ data without sacrificing privacy.
This is also being used in the financial world. I am working on utilizing federated learning for financial crimes. Banks don’t want to share information with each other, but with federated learning, they are able to find out if certain accounts were involved in money laundering and we can come up with a model to predict if financial fraud was committed.
Those are just a few examples. It is also being used in transportation with self-driving cars. It is used by Google, Apple, and Facebook with text predictions. All of these companies cannot use personal information in text messages due to privacy, but they can use the data to train models to make the experience better.
What courses are you teaching this fall? What can students expect to get out of that class?
I am teaching a grad-level research class - CSCI 8980: Topics in Modern Distributed Systems. I will be focusing on advanced distributed systems that enable modern applications, federated learning, server-less computing, and cloud-storage. This class will focus on teaching students how to design a system from the ground up for a large-scale application. I have a couple of possible options outlined so students can pick an area that is interesting to them.
I am actively looking for students for my research labs - undergrad, masters, Ph.D. - all are welcome. They can reach out to me and I will help them build a plan that is specific to their goals. With undergrads, I especially take the time to figure out what they want to do after graduation - if they want to go into the industry or grad school. The plan is centered around the students. For Ph.D. students, their experience would be heavily research focused, and for the masters students it would be in between the undergrad and Ph.D. focus.
What do you do outside of the classroom for fun?
I have a lot of hobbies. I like painting and going for runs. I also play a lot of video games and read a lot. I’m really into fiction and I have started some of my own creative writing, but it is not very good yet. I enjoy both science fiction and fantasy. Recently I have been enjoying Andy Weir who wrote “The Martian”. I have read the Lord of the Rings, Harry Potter and the Song of Ice and Fire series. I also spend a lot of time with my family.
Do you have a favorite spot in the city?
I really like Afro Deli downtown. I also enjoy being downtown when there are sporting events. It is a really fun atmosphere.