Steven S. Lumetta
Mark Pajor, ECE ILLINOIS
Q: What is your area of expertise?
A: My general area is computer systems. I did high-performance computing as a graduate student. Once I came here, I started working on computer architecture, mostly designing processors. I also worked in optimal network architecture, which was looking mostly at failure recovery algorithms and how they integrated with the changes in the hardware at the time. I’ve worked in digital system testing a little bit, so we’ve looked at using coding theory to do compression of test vectors coming out after they tested each of the test vectors in the chip. I’ve used the high-performance computing side to help with the Blue Waters supercomputer proposal a few years ago. I was one of five or six faculty originally involved in that here at Illinois.
More recently, I’m one of the leaders in our computational genomics effort, and am the principal investigator for one of the main grants. The effort was started by Gene E. Robinson and Ravishankar K. Iyer, and I wrote what’s called a major research infrastructure grant. Gene, Ravi, Victor Jongeneel, and Saurabh Sinha were my co-PIs. Now we have 35 faculty involved in that grant and about 50 in the bigger initiative. It’s gathering momentum on campus, so it’s pretty exciting.
Q: Can you give a brief synopsis of your education and career?
A: I was an undergraduate at Berkeley in Physics and English, graduating in 1991. I was an experimental astrophysicist as an undergrad, and then I went into computer science graduate school, also at Berkeley. I earned my Master’s in 1994 and PhD in 1998, and then came straight here to Illinois. As for industry experience, Muriel Médard and I worked on optical networks together. She, her husband, and I started a company in 2000. That ran for about four years and got up to about 75 people.
Q: What do you enjoy most about being at Illinois?
A: The environment with the faculty and the students is the biggest attraction, having a lot of smart people around hanging out and talking about problems. That’s the most fun. And the community is nice. I grew up in Berkeley, and it’s nice to know neighbors here. Letting my kids play outside and feeling safe about it is nice. The public schools are good, and there are good private school options. There are good school options generally, which is not true in every state.
Q: What do you enjoy about teaching?
A: I think my favorite part is interacting with students. That’s the most fun, when the students are excited and realizing that they actually know a lot. I think that because everyone here comes out of the top of their class, when they come here, they first they feel a little depressed because half of them are under average. I tell that to my advisees, and then I let them know that half of grad students are also under average, and if you decide to stay in academia and become a faculty member, it turns out that half of faculty are under average, too. The important part is to recognize the peer group. And then when our students go out and do internships, they realize that they actually know quite a bit. When they get to know students from other schools, they feel more proud. So it’s nice when the students are getting excited about material, and enjoying things, and feeling like they’re learning a lot.
Q: Which of your research projects are you particularly excited about?
A: I’m excited about all the things I’m working on, but the bioassay project that Brian T. Cunningham and I have is pretty cool. The technology came out of his lab, and I’m doing the app side of things. The idea is to have a cradle in which you can put a cell phone. The cradle has a grating over the camera so that you see a spectral response. We’ll use the camera flash LED and a laser pointer as light sources, and will be able to do several common biological assays. I believe Brian’s lab has prototyped all of them at this point, and published papers on the ability to get good results out of those kinds of assays. We’re making an Android app so you can have it in the field. We have classes doing biological experiments and bioassays, and they will be able to use this as lab equipment.
But probably more importantly, around the world many people have cellphones, but that’s about the limit of their equipment. Now you can use this in villages in most of the world, reach the other billions of people who don’t have anything else in terms of infrastructure, and do things like genomics tests and epidemic control. There are all kinds of cool applications. The phones are already there, and if they’re not, they’re cheap. There’s a company that does something else with a $50,000 piece of equipment that we can replace with something like this same cradle and a light bulb. So it’s a cool technology.
Q: What does the future hold?
A: I think of computational genomics as the exciting science of the future for bright young people, in the sense that there’s always something that’s moving science research, having impact on the world. The reason I got excited about it is that the cost of sequencing has come down so much that we’re really starting to see a lot of data. And when you move from a data-poor environment to a data-rich environment, it becomes a different set of problems. You can make a lot more progress much more rapidly. We had a workshop a couple of weeks ago. There was someone who had been working in the area for a long time, and he told us at dinner: “When I was in grad school, you would be happy if in the space of a week you’d gotten 100 base pairs out of your sequencing.” And these days, the number of genomic data is doubling basically every five months.
Here’s a way to think of that: one project in the last year has gathered as much data as all previous projects in the previous 20 years. There was a 1,000 human genomes project that got to 1,000 well before it finished, so they decided to go for 10,000. The people at BGI, the Beijing Genomics Institute, are talking about a million genomes project. The numbers are ramping up and up. We will have plenty of data, and the question is – how do we use those data effectively to move the science forward? And I think it’s an exciting time for that. So I think over the next year we’ll see a lot of changes. The grand challenge is matching genotype to phenotype, so with all this data, I’m hoping we’ll be ready to start doing models, and understanding the connections between the genome and how it operates in practice. If we could start to build better models that allow you to make better connections and predictions between genotype and phenotype, that would be pretty cool. I think we’re on the cusp of that revolution.
Q: What do you hope to accomplish with your research?
A: With computational genomics, I hope both to have some research impact myself, as well as really enable Illinois to lead the way. The infrastructure grant really creates this resource that the whole campus can use. It means that we have a machine where if people want to dabble with new ideas that involve interdisciplinary groups, they don’t have to get funding first. There’s a machine, and they can say, “Hey, we have some interesting ideas. Can we play with this?” And we can let them play with it. We can use their results to try to design the machine; it’s really a machine building project, as well. Closing that loop and giving this resource to the campus is pretty important, and I’d like to see that bear fruit.
I’d like to see a lot more efforts like that on campus. As for my other projects, I’d like to see them succeed and be used widely. So I think that the answer to your question is really, “impact.” And there are a lot of ways for that to happen. Like I mentioned, I‘ve started a company in the past, and lots of ECE faculty members have started companies. That’s one way to have an impact. And handing it off to Intel, AMD, IBM, Microsoft, NVIDIA, whoever wants to pick it up, that’s another way. And you can send off students who will build startups or join industry. So I think there are many ways for that impact to happen, and I think that’s the main goal for most of us in ECE: to change the world.