The University of Toronto’s Student Newspaper Since 1880

Share on facebook
Share on twitter
Share on email

U of T students place second at IBM Watson University Competition

Developers behind ROSS on entrepreneurship, education, and the future of technology
Share on facebook
Share on twitter
Share on email
Akash Venkat, Andrew Arruda, Shuai Wang, and Jimoh Ovbiagele (from left to right). NATHAN CHAN/THE VARSITY
Akash Venkat, Andrew Arruda, Shuai Wang, and Jimoh Ovbiagele (from left to right). NATHAN CHAN/THE VARSITY

A group of U of T computer science students recently placed second in the IBM Watson University Competition, presenting their app, ROSS, that functions as a legal assistant.

The app — based on IBM’s artificial intelligence platform, Watson —  shortens research time for lawyers by analyzing case files and performing legal research using artificial intelligence. ROSS competed against top American schools at the culminating challenge in New York City, after besting four other U of T teams.

The Varsity spoke with Jimoh Ovbiagele, Shuai Wang, and Akash Venkat who created the app along with teammates Pargles Wenz Dall’Oglio and Andrew Arruda.

The Varsity: Tell me a little bit about yourselves and how you got involved in building ROSS.

Jimoh Ovbiagele: I got an email from the undergraduate newsletter and I just picked up the words IBM Watson. So I clicked on the link and it pulled up on the computer science forums, which was… really dated… But in there was the gem: the opportunity to work with Watson… So I applied, sent in a résumé, they asked some critical thinking questions, and then I got an email telling me that I was accepted.

Shuai Wang: For me, I’ve always been interested in computer science courses, so I looked through the website of computer science, and I tried to select a bunch of courses for the semester, and I got hooked by this course.

Akash Venkat: It was two courses, kind of offered under one umbrella, so it was traditionally the Business of Software class, but they had this cognitive computing capstone added in because of the Watson contest. That was really it. For me, I was just kind of going through cool CS courses to take and then I saw that they had this Watson component, and I’m like, this is fantastic.

TV: What was the competition like?

AV: It was amazing; just being in the IBM Watson building was pretty surreal, I think, for all of us. I mean, this is kind of the place where they work on Watson every day and it’s where they built the Jeopardy winning machine, and I think the biggest thing was just interacting with the executives running the company, sort of that close.

So we were presenting to the four or five people actually running the IBM Watson group, and so it’s just that proximity to people that important [that] was pretty sweet. The field of schools that actually competed was the top CS departments in North America, so I think that really raises the stakes of the contest.

They kind of took us on a tour of the actual facility. It was like a client experience to us. They showed us what Watson’s about and how they impress Watson clients, essentially.

TV: What was the most memorable part of the project, and what was the hardest part of the competition?

AV: I think we really, really enjoyed it from the get go. The good thing about our team was that we really gelled very well together from the start. And it was just the mindset as well. We thought of this as an actual company… an actual startup really early on, and that shaped the culture of how we work together. I think that made a big difference.

JO: It’s hard to say what’s memorable cause we’re working in the moment, but we’re thinking ahead so we don’t really have time to think back on things. New York felt like months ago.

AV: Yeah, exactly, I mean the two days was like a blur, it all went by so fast. We were all focused on the presentations and even the social stuff after; all that just went by really fast. I guess the really memorable moment was since we did so well, I think the next day morning, when it kind of sank in a little, I think that was pretty special, is when we knew our lives had kind of changed at that point. So that was the moment we had to reflect on what we had done, so I think that was a pretty special moment.

It’s been full of memorable moments. It’s been amazing wins, from the first bit of coverage we got in the Metro to the… projects we’ve signed with law firms, and just going forward now, we’re doing this full-time.

TV: I suppose you had a lot of options for what to do with Watson. Of all the options you had, what inspired you to make a legal researcher?

JO: Well, a lot of the big domains where Watson could make a big impact, IBM already had their hands in, like medicine and finance. The next big thing was law, and it’s an industry that’s sort of resisted technological change. So we saw a great opportunity to make the legal system better and provide better legal services to people like you and I, but also corporations and governments.

AV: Just the nature of the data. So it’s a lot of unstructured data, and that’s what Watson’s really good at mining conclusions from, so it felt like a great fit.

TV: Can you explain in nonscientific terms how Watson works?

JO: So, what Watson does is it breaks down sentences, and it understands the relationships between words. From that, it relates those relationships to the relationships… found in other sentences, so it knows how a word is used in the context of a sentence but also in the context of some subject matter. So… if I say a word, it really means nothing standing by itself. It only means something in a sentence, and that sentence only means something because of your experiences and your previous knowledge that inform how you interpret that sentence. So that is really the big thing that Watson is able to do, is to understand the context of a sentence and being able to understand humans in ordinary language as opposed [to] something like keyword search, which we are all used to doing. Since the advent of computing, we’ve been used to, or forced to, communicate with computer[s] on their terms, and this is turning that around.

SW: So, there’s a vivid example. For keyword search, if you type in Google, “find me any restaurant but a pizza restaurant,” all you get is a list of pizza restaurants. What Watson does here is generate hypothesis and does some semantic and syntactical parsing so it [can] analyze the content like Jimoh says. That’s the main difference between natural language processing and the keyword search.

TV: Entrepreneurship is a big leap from, you know, going to school. How did you get into it?

JO: I feel like the answer is going to be different for each of us. At least for me, school and education was always a means to an end. So, the destination was always tacking some challenge, which entrepreneurship, I think, is. So, now, leaving school early to do this, it was for me just a realization that I could tackle this big problem now.

AV: I think I was always really interested in entrepreneurship, so I had my first little taste of it when I worked at a start-up as an intern, and that was really addictive, I kind of got the whole start-up bug during my experience. So I knew that was something that I wanted to do at some point. It was a question of the right opportunity, and the right team, and the right market, and the right product. And all of that just came together beautifully during this project. And I think that a lot of it was the mindset. So the reason we did that well was because we considered it as a serious entrepreneurial gig, and that’s kind of what raised the stakes for each of us. And I think that really helps even in education. You kind of think of things being on the line. You know you have to work that much harder. There’s more at stake.

TV: How did your U of T education help you with the project, both from a technical point of view and an entrepreneurial point of view?

JO: …I don’t think, because of the resources that are out there, the education is unique. What I do think is [unique, is] the people that you meet here and the exchange of ideas. Because, if I had never met these guys here, I don’t know where else I would have met them. So, yeah, I think that U of T is a meeting of minds. But, in terms of the education, which I know that a lot of people are expecting to hear, we are at a great computer science department so we’ve attracted many great computer science minds, and minds in the information school as well but, it’s more the people.

SW: …The reason why that this course could give us such a good experience is that it combines industry and academics. The education institution shouldn’t be the Babel Tower. We should expect more interaction between industry and real academics.

JO: The big theme here is engagement. It’s not unique about U of T. It’s a characteristic of the current education system from kindergarten to now: it’s not engaging. It lacks the why are we doing this, why are we here?

TV: How do you see the role of computer technology changing in the next ten years?

JO: We’re at the beginning of a new industrial revolution. The original industrial revolution mechanized muscle and now with these new computing advances like machine learning, deep learning, natural language processing, we’re automating knowledge work. So, there’s going to be a lot of things that change in that aspect. I think we’ll be a more productive society, and it’s going to accelerate our investigation into other aspects than computing. For the first time, we are augmenting our ability to think with machines… And another thing where I think technology is moving: technology becoming invisible because technology is becoming more intelligent so it could disappear when it’s not needed and reappear when it is.

TV: So, what’s next for Ross?

JO: Our vision is to build a great Canadian company… What we’re doing right now is, we’re putting the systems in place so that we can execute on that mission. We’re forming partnerships with IBM and we just formed a partnership with one of the largest international law firms in the world. So, getting those partnerships in place. Also, making key hires, seeing what skills we cover and then seeing what skills we need, and then hiring the difference. And also, raising money to fund this. The interest to get behind this from an investment standpoint has been amazing. That’s what we’re doing. We’re preparing for launch.

TV: Going into entrepreneurship… isn’t it a little bit scary? Tell me a bit about your experience — was it exciting, scary? How do you feel about all this?

AV: Yeah it’s definitely scary but what I’m really confident about, other than the fact that we have a really phenomenal team and we work really well together is the fact that the time for the product is right. The market needs something like this — it’s really screaming for a solution like this, and we have the cure. So we’re bringing that to the table. We’ve had validation of that really early on. It’s not like we’re building a social app and hoping to grow a massive consumer base and then figure out a way to monetize that going forward. We have all that in place already, and we have an amazing team of advisors supporting us, and obviously IBM watching our back as well. Having that entire support system in place this early is really reassuring, so, while it is scary in that we’re embarking on something that none of us have actually done at this scale before, we couldn’t ask for a better position to do that from.

SW: And it is also exciting. I mean, we have so much to learn and we have something that can wake us up in the morning. There’s something really passionate about: it’s just amazing.

TV: What’s next for you guys personally?

Everyone: Ross.

Correction (January 27, 2015, 2:43 pm): A previous version of this article incorrectly stated that ROSS predicts outcomes for legal cases. In fact, ROSS performs legal research using artificial intelligence. The Varsity regrets the error.