Who is Professor Matthew Taylor? | What it's Like to Be

Read the Professor Matthew Taylor section.Professor Matthew Taylor#

Read the Drones, Atari, and Artificial Intelligence section.Drones, Atari, and Artificial Intelligence#

Professor Matthew Taylor is a Professor in the Faculty of Computing Science at the University of Alberta. He specializes in artificial intelligence and reinforcement learning.

He teaches CMPUT 412 and 503 at the University of Alberta.

Check out his personal website over here: https://drmatttaylor.net/

You can find out more about his research here: https://apps.ualberta.ca/directory/person/mtaylor3

Check out his research publications here: https://scholar.google.com/citations?user=edQgLXcAAAAJ&hl=en

This interview was really fantastic, and Professor Taylor dropped a lot of useful knowledge. Reading through all of the transcript might be difficult, but it will be worth your time.

Read the Highlights section.Highlights#

Here are some important tidbits from the interview:

Check out the dialogue at 4:10. Professor Taylor discusses networking.

Read the Matthew Taylor 4:10 section.Matthew Taylor 4:10#

Well, I’ll say that in computing science. Being an introvert is not all that abnormal. And it’s just a matter of, so for me, one of the hardest things to do is go to a professional conference and network with people for like three days straight. And I tried to do that a number of times, and I just exhausted myself.

So now I know when I go to one of these conferences, I need to spend time back in my hotel room on my own, just recuperating. Yeah, you know, and part of me just wants to spend the entire conference hiding in my room. I can’t do that. But you compromise and figure out okay, I want to try to meet three new people today. You know, you set that kind of goal and then if you don’t make it, that’s okay. But then when you do make it, you kind of congratulate yourself Okay, and I have friends who would go out and meet 20 people. Yeah. But I just I realised that I, I’m not that person. I’m going to work with what I have.

Check out the dialogue at 6:33. Professor Taylor discusses his favourite projects.

Read the Matthew Taylor 6:33 section.Matthew Taylor 6:33#

Let’s see. Okay, so in one of the coolest projects I had, we used drones to chase birds away from high value crops. So you think you’ve got a vineyard, you’re growing these expensive wine grapes, and the birds come in and eat them. The state of the art method is you give guns to a bunch of teenagers, and say, go out and shoot or scare the birds.

In this project, the thinking was let’s do something a little more humane and we can just use these unmanned aerial vehicles to go scare the birds away. And then you can use this for airports or beaches or other things. It’s thinking about, okay, we’ve done this thing the same way for years. Now we’ve got this new technology, is there some way we can leverage that to try to beat this old problem with new techniques?

Check out the dialogue at 18:48. Professor Taylor discusses some cool Startups in Alberta.

Read the Matthew Taylor 18:48 section.Matthew Taylor 18:48#

Well, med.io is pretty cool. They’re doing some cool imaging stuff. There’s also a cool Startup, a medium small-sized company in Calgary called Attabotics. They have these robots that work in warehouses. This robot goes, grabs your socks, brings it to someone and then puts it in the box, and then boom, you get mailed your socks. You don’t necessarily hear about these homegrown companies, because we don’t, it seems like Canadians might not be as good at trying to show the world that look, we’re doing all this cool stuff. I think this is a stereotype, but I think Americans are more likely to be saying “look at me, look at how awesome I am!”

Check out the dialogue at 24:00. Professor Taylor discusses Deep Q Networks.

Read the Matthew Taylor 24:13 section.Matthew Taylor 24:13#

Yeah. So this is the same type of algorithm that worked for AlphaGo, a program that beat Lee Sedol. And the idea is that you could have a neural network, which is just again, an abstraction. You could have this neural network and try to figure out what was important. So if the neural network is looking at an Atari screen, maybe over time, it can figure out oh, this is the player, I should pay attention to the player. Or if it’s looking at a goal board, oh, this is a certain type of block of stones that I need to pay attention to. So in the past, it was all up it was mostly up to the the nerd the programmer, to say, here’s what’s important. And with this deep UL algorithm now the algorithm itself could figure out what’s important.

Check out the dialogue at 28:06. Professor Taylor discusses his most difficult challenge.

Read the Matthew Taylor 28:06 section.Matthew Taylor 28:06#

Okay, well, this is easy. I was a software developer and I applied to 11 graduate schools. I got into one. Okay, so clearly, they made a mistake, but they let me in. And then I found this advisor, he was doing amazing stuff. And he said, okay, you’re going to do that.

If we found a project, I would work on it. And it was a pure C coding project. And I jumped in. Turns out, it wasn’t that good at coding and C. So after, after a number of months, I had made almost no progress. I was I was ready to drop out of grad school. And then I talked to my professor again, you know, clearly this wasn’t going well.

So we found a different project that involved less coding, but slightly more thinking. And thankfully, I succeeded there. So in that case, initially, we had a real mismatch between what the project required and what I was capable of. And at the time, I didn’t know that and I just thought I was stupid. Right?

So I’m really down on myself. But then luckily, we found another project, which was better aligned. And then things went well for the rest of grad school.

Read the Full Interview: section.Full Interview:#

Read the Matthew Taylor 0:29 section.Matthew Taylor 0:29#

Well, and at my previous school, one of my best research students was an undergraduate who just walked by my office and said, “Excuse me, what research do you do?” And we clicked and we worked together a lot, but most students don’t realise that this is something you can do, or most students don’t realise that there’s all this stuff happening. You just need to find out about it.

Read the Abdul-Samad Olagunju 0:50 section.Abdul-Samad Olagunju 0:50#

Yeah. I think that even in my case, that applies to me. I’m in third year now. But in my first year, I did not do any research of any kind. And I think, after a while, when you start learning about stuff in class, a part of you just wants to apply it and see how it works. And I think, yeah, research is a great way to learn more and apply your knowledge. But enough about what I’m saying. This interview is obviously about you.

I learned a bit about you from your website, and the fact that you specialise in machine learning really grabbed my attention. Anytime I see anything about machine learning or AI, I’m already hooked. So I thought, okay, if we can get this thing going, I surely interested in learning about what you do. My first question is just like, how did you get here? What was your educational path that led you to this position?

Read the Matthew Taylor 1:42 section.Matthew Taylor 1:42#

Yeah, so I was an Undergraduate, and I wanted to do Physics research, but that was too hard. So I started doing some Computing Science research. And that was too boring. So I became a software developer. And then that was boring. So then I went back to grad school and what’s the coolest thing I could do? Well, there’s this professor that was doing stuff with robot dogs.

It’s like, yes. Okay, robot dog soccer. And that’s awesome. So, you know, this was around 2003. And I just, AI was the most fun thing I could think of doing. And then since then, I’ve kind of bounced around the US to a bunch of places. And then the kind of AI, the type of machine learning I was doing actually became useful. So it wasn’t just playing computer games. But I got hired by the Royal Bank of Canada, to come up to Edmonton and help start an AI Lab. Because our machine learning could actually make money and do stuff for people.

Read the Abdul-Samad Olagunju 2:41 section.Abdul-Samad Olagunju 2:41#

So how do you establish those kinds of connections? You said, you’ve moved around a lot?

Read the Matthew Taylor section.Matthew Taylor#

Yep.

Read the Abdul-Samad Olagunju section.Abdul-Samad Olagunju#

So when you move around, you’re going to have to change your organisational base to establish it somewhere else? How do you take that risk and make that jump? Can you explain that further?

Read the Matthew Taylor 2:56 section.Matthew Taylor 2:56#

Yeah. I mean, partly, you’re absolutely right. It is a risk whenever you move. But I think when you when you look around, like where are the best opportunities for me, that’s often not where you currently are. But then you go, you go, you move across the country, you move to a different country. And you still have all those same connections via email, via Google, Google Chat, all of that.

But then you’re right, you want to meet people in your new environment. And for an introvert, like myself, that’s really hard. But, you know, you put yourself out there, you start making friends. And because now I’m getting more senior, when someone comes to me, I can help make that introduction, someone I know. And vice versa. When I meet someone more senior than me, they know people can say, hey Matt, you need to go meet this guy. Because it’s so much easier if someone else can say, you know, based on what you know, based on what you’re interested in, you need to go talk with her.

Read the Abdul-Samad Olagunju 3:58 section.Abdul-Samad Olagunju 3:58#

Okay. And you say, I’m an introvert. So how did that affect your rise in academia? You know, getting this position?

Read the Matthew Taylor 4:10 section.Matthew Taylor 4:10#

Read the Abdul-Samad Olagunju 5:09 section.Abdul-Samad Olagunju 5:09#

Yeah, I think that’s also important that people recognise they have different limitations. And you have to maximise what you’re able to do. Yeah, I think over the pandemic, for me personally, all these kinds of ideas started coming in my head, like, oh, yeah, the best way you should go about doing something is so that you can try your best at it. But also, you don’t over exhaust yourself.

So that moves me on to another question. How do you manage all the workload? You know, it’s a crazy amount of work as a professor, you have your own lab, you have all your research projects going on? How do you manage all of that?

Read the Matthew Taylor 5:46 section.Matthew Taylor 5:46#

Yes, the problem is, as a professor, there are all these cool things you could do. And it’s kind of like going to a all you can eat buffet. And you know, there’s so many things. And you realise if you take too much food, you’re going to end up regretting it later.

So the trick is trying to figure out what you can commit to. And then once you once you have those set of commandments, then it’s a matter of balancing and prioritising and trying to figure out, you know, mate, maybe I’m really good at this thing. But this other thing I tried, someone else could probably do that.

Read the Abdul-Samad Olagunju 6:22 section.Abdul-Samad Olagunju 6:22#

Yeah. So what’s the most interesting project that you think you’ve had? I know, it’s a difficult question, but just in your opinion…

Read the Matthew Taylor 6:33 section.Matthew Taylor 6:33#

Read the Abdul-Samad Olagunju 7:24 section.Abdul-Samad Olagunju 7:24#

Were these drones, like, was someone controlling them? Or were they autonomous?

Read the Matthew Taylor 7:33 section.Matthew Taylor 7:33#

They could be both. So you could have a drone just fly a pattern, you know, constantly, but we didn’t get quite this far. It’s better if you can automatically detect the birds, fly towards the birds, and scare them away. Because if you have to pay someone to operate drones that’s much more expensive.

Read the Abdul-Samad Olagunju 7:49 section.Abdul-Samad Olagunju 7:49#

Yeah. Because this summer, I’ve been getting more interested in Python, and also web development. So when you start designing these algorithms for the drone, it’s probably got a camera on it, it sees the bird, it identifies it. So how do you make sure that it knows to classify that as a bird? Are you using external programs? Are you writing your own?

Read the Matthew Taylor 8:13 section.Matthew Taylor 8:13#

Definitely using external libraries? In machine learning, if you started building from the ground up, you’d never get there. You want to steal and borrow as much as you can. So in our case, we, we stole some existing vision libraries. And we assumed that anything that’s a kind of above the horizon is a bird that’s moving.

So if it’s a plane way off in the distance, we just got it wrong. But then you can try to see if something’s moving in a straight line, okay, I’m going to assume that’s a bird so I can move towards it. And then, it is actually not all that complicated. Because you just figure out which direction is the bird in, and then I go forward in that direction.

Read the Abdul-Samad Olagunju 9:01 section.Abdul-Samad Olagunju 9:01#

And so you’re going to write your programs on your computer? How do you share your code, get everyone on the team in the same direction, in the same mindset, so that you can accomplish this project on time?

Read the Matthew Taylor 9:19 section.Matthew Taylor 9:19#

Yeah, that’s that can be tough. Especially in school, because you have lots of different people with different abilities and strengths. And there’s really figuring out who can do what, and trying to figure out that you want one person who is just going to work on identifying the bird. Another person is just going to work on taking off and landing safely. And then it’s a matter of, you know, integrating all of this so that hopefully all these different modules play nicely.

Read the Abdul-Samad Olagunju 9:47 section.Abdul-Samad Olagunju 9:47#

Okay. And what’s your preferred leadership style? Do you like to micromanage or are you more of a hands off kind of guy?

Read the Matthew Taylor 9:55 section.Matthew Taylor 9:55#

I am personally much more hands off. I mean, for me, the big thing is I want to recruit students into my lab and figure out what they want to do and how they can succeed. My job is to try to come up with an environment where you come in with your skill set and what you’re excited about. And I help you figure out what you want to do and then accomplish it.

Read the Abdul-Samad Olagunju 10:30 section.Abdul-Samad Olagunju 10:30#

So what kind of student do you think is a good fit in your lab? What attributes are you looking for in a student?

Read the Matthew Taylor 10:37 section.Matthew Taylor 10:37#

Yeah, good question. I guess for me, students have to have some background in machine learning and AI, like you know, one undergrad class. And then it’s more of, a lot of what we do in AI, we don’t know what we’re doing. And it’s finding people who are comfortable with that—I don’t want to say chaos, but are comfortable with that uncertainty that, you know, we don’t necessarily know what we’re doing on Monday. But hopefully, by Wednesday, we’ll figure out what the right direction is. And then if it’s wrong, we’ll try something else.

Read the Abdul-Samad Olagunju 11:15 section.Abdul-Samad Olagunju 11:15#

And, again, you’re talking about the chaos in this industry of AI? How do you keep up with it? You’ve been in the AI for 20 years. How do you keep up with all the technological advances all the new things that are happening?

Read the Matthew Taylor 11:31 section.Matthew Taylor 11:31#

It’s so hard, and it’s getting faster and faster. And if you just try to look for new papers in your subfield, you get like five new papers a day, there’s no way you can read that. So a lot of what I do is rely on my students and colleagues to say, hey Matt, have you seen this? And at the same time I try to read one or two papers a week. And after I read that paper, I think, okay, who would want to know this? So you kind of build that community where you try to share knowledge because no one person can possibly keep up with all this?

Read the Abdul-Samad Olagunju 12:07 section.Abdul-Samad Olagunju 12:07#

And I don’t know, maybe this is my personal opinion, do you feel like scientific research should be more accessible? I don’t know, when I first started reading papers, this is in neuroscience, because I’m in neuroscience, It just felt like tough with all of the small font and everything packed tightly together, and a lot of highly condensed information.

I still think we should have articles like that. But do you think we should also care more about young scientists and maybe people in the public who want to learn about this stuff, maybe have an easier kind of manuscript for them to read to learn about?

Read the Matthew Taylor 12:42 section.Matthew Taylor 12:42#

Yes, absolutely. There is some online experimentation. There’s some online journals that focus more on good videos and animations. But the problem with science is, you’re often incentivized to come up with the newest, newest thing. And once someone has built up a lot of these new things, it’s really useful if you can take a step back and say, okay, how do all these things relate? And how can I package this in a way that a non expert could understand? And there’s not, I don’t think there’s enough emphasis on that. Because if you’re, if you’re a neuroscientist, there’s lots of cool stuff in AI that you could find interesting and learn about. But it helps if you’re given that high level picture first. And most of the time, you don’t have to delve into those tiny print journal articles. You know, because it’s the high-level ideas that could be really useful. Not all the tiny details that you need to implement this particular neural network.

Read the Abdul-Samad Olagunju 13:47 section.Abdul-Samad Olagunju 13:47#

So when you you’re about to start a project, you’re not focusing on the details, you think of the broad idea, you think about what technology you want to use. How do you narrow in on, okay, I want to use this to learn about this? How do you narrow in on that thing that you’re going to use for your project?

Read the Matthew Taylor 14:08 section.Matthew Taylor 14:08#

Well, part of this goes back to how people conduct research. One direction is saying I’ve got this new idea, I built this hammer. And now I’m going to find the best problem that I can use this hammer for. That’s a completely legitimate way of doing research. Another way is saying, I found a problem that’s important. I’m going to find the right tools or build the right tools to solve this problem. And that’s kind of what I think you were thinking about. Once you’ve identified a problem, how the heck do you figure out how to solve it? And there for me, it really goes back to who in your network might know.

I mean, I can think about what kind of tools I might use. But then let me call up Cynthia and see what she thinks because I know she did something that was a little bit related five years ago, maybe she has some ideas. And just talking to people can be just so incredibly useful. Because, you know, the three things you think of may not be nearly as good as the 30 things that your friends tell you about.

Read the Abdul-Samad Olagunju 15:19 section.Abdul-Samad Olagunju 15:19#

Okay, so do you feel like it took you time to learn this? Or did you know this at the start of your career?

Read the Matthew Taylor 15:25 section.Matthew Taylor 15:25#

It definitely took me time to learn. You know, I always thought science was about sitting in a room and thinking hard. And if you’re just smart enough and try hard enough, you’ll fix it. You’ll come up with the answer. And I’ve realised that there are so many people that know so much and are so smart. Yeah, it’s kind of silly for me to just try to do it all on my own.

Read the Abdul-Samad Olagunju 15:50 section.Abdul-Samad Olagunju 15:50#

Yeah. So it’s an interesting thing. Do you think there is an argument against the Nobel Prize? In an article, some person was saying, you know, science is a collective thing. Even if someone makes a discovery, hundreds of thousands of other people have helped them get to that point. So we should appreciate the collective, not the individual. So when you’re working with other people, and you come up with your own idea, is there occasionally some conflict if someone doesn’t support your idea? How are the dynamics of the situation?

Read the Matthew Taylor 16:31 section.Matthew Taylor 16:31#

Yes. So normally, people come up with lots of ideas, and are biased towards the idea that they thought of, you know, because if you come up with it, you’re more invested. I mean, assuming that you’ve got people that are excited about the science, and not just self promotion, hopefully, they’ll realise okay, maybe this one idea is better than the other, or we don’t know which is better. So let’s figure out how to test both of them in a way that we can agree on which one would be better going forward after we do this test.

Read the Abdul-Samad Olagunju 17:05 section.Abdul-Samad Olagunju 17:05#

So it’s more of a collaborative environment, everyone’s working together?

Read the Matthew Taylor 17:11 section.Matthew Taylor 17:11#

Right. And if the two of us are sitting, we’re sitting around this table, and I say something and you say something, it doesn’t really matter who said what, as long as we can say, well, we came up with this idea together. And then we can show whether it works or not.

Read the Abdul-Samad Olagunju 17:29 section.Abdul-Samad Olagunju 17:29#

Yeah. And you said that you spent some time in the States as well. So how is it different to Canada? The technology industry, the people in academia? How is it different to what you’ve seen over here?

Read the Matthew Taylor 17:43 section.Matthew Taylor 17:43#

People are definitely friendly over here. That’s a big one. Other than that, I don’t see huge differences in terms of academia or tech. I mean, one thing Alberta hasn’t done great in the past is getting our students to graduate and stay here. Oh, I think we’re doing better now.

But a lot that, you know, there’s a big incentive in Computing Science, you graduate, and then go to San Francisco and make a lot of money. Right. And it’s better for us, Canadians, if we can get students to succeed in school, and then stay here, and then work in Canadian companies or Albertan companies. Being in the US, it was very common that people would graduate and go to the Bay Area, or to New York City, and make lots of money. And it was, it was less common to kind of stick around where you got educated.

Read the Abdul-Samad Olagunju 18:41 section.Abdul-Samad Olagunju 18:41#

Are there any special Startups right now in Edmonton that you’ve had your eye on?

Read the Matthew Taylor 18:48 section.Matthew Taylor 18:48#

Read the Abdul-Samad Olagunju 19:40 section.Abdul-Samad Olagunju 19:40#

The pride. Yeah, I think that’s really interesting. Pivoting a bit more to the University of Alberta. Why did you decide that the University of Alberta was the right place for you to?

Read the Matthew Taylor 19:57 section.Matthew Taylor 19:57#

In my field, Rich Sutton is kind of one of the founders of the field. And Rich is a professor here in AI. So Rich was on my PhD committee back in 2008. And like, since then I’ve been looking for an excuse to get up here. For me, it was very easy, because in my subfield of reinforcement learning, this is easily either the best school or certainly in the top three in the world. So I mean, for me, being able to, you know, be down the hallway for someone that I consider a legend is pretty amazing.

Read the Abdul-Samad Olagunju 20:34 section.Abdul-Samad Olagunju 20:34#

Yeah, I didn’t even know.

Read the Matthew Taylor 20:37 section.Matthew Taylor 20:37#

And that’s because we don’t yell enough about how awesome we are.

Read the Abdul-Samad Olagunju 20:41 section.Abdul-Samad Olagunju 20:41#

Exactly. And in reinforcement learning, you know, I don’t know anything about AI, the only machine learning I’ve done is maybe a playing around with a couple functions in Python. And then you run a data set and get some outputs.

But how do you keep track of, okay, this is the right algorithm for this situation versus this algorithm? Are you testing a whole bunch of algorithms at once before you find the right one? Like, how does it work, when you’re not just like a casual person sitting on your computer playing with what someone’s built, but like, you’re actually a research professor, creating projects and doing all this?

Read the Matthew Taylor 21:16 section.Matthew Taylor 21:16#

So there’s, again, kinds of different ways of doing research, one would be, I’m going to grab any reinforcement learning algorithm, and try this new thing and show how I can make reinforcement learning better in general.

In that case, you can grab an algorithm for 20 years ago, and it’s fine. But then there’s also the case where I’ve got this specific problem. And I want to go from 90% to 91%. And in that case, you’re going to try out 20 algorithms. And each of those algorithms is going to have a few knobs. So I’m going to try turning those knobs.

And eventually, by using lots and lots of compute, we’ll find a good combination of algorithm and settings. And we’ll do really well at this one problem we really care about.

Read the Abdul-Samad Olagunju 22:03 section.Abdul-Samad Olagunju 22:03#

Okay. And with these algorithms that you’re getting, does it matter that you understand it, like fully? What the computer is doing, how much memory it’s using?

Or what variables do you take into account as you’re thinking about, okay, this algorithm, even if it’s a higher accuracy, are you thinking about memory storage, and all these other things?

Read the Matthew Taylor 22:27 section.Matthew Taylor 22:27#

No. And one of the key concepts of Computing Science is abstraction.

So I can think at a very high level, like, what is my percentage accuracy? I could think at a medium level, like, what is the architecture of my neural network and the memory usage? Or it could even think at a very low level. So think about what are the different gates that I’m using inside my microprocessor. And normally, you would not want to go that low, because we’re not electrical engineers.

But you’re right, that depending on the problem, you might want to think at different levels of abstraction. And in some cases, I don’t care about memory, I just want the thing to work. But if I want the thing to work on a cell phone, and that’s memory constrained, I better think very carefully about what kind of memory I’m using.

Read the Abdul-Samad Olagunju 23:19 section.Abdul-Samad Olagunju 23:19#

Okay. And what is the name of that your favourite algorithm, the algorithm that you used in a project and it just made you so happy, or it works beautifully?

Read the Matthew Taylor 23:32 section.Matthew Taylor 23:32#

Yeah. So I think my favourite algorithm is called DQN, Deep Q networks. And this was the algorithm that let computers play Atari Games by just looking at pixels. So, the computer can just look at the pixels that the Atari game is showing, and figure out how to move the joystick to get lots of points. And this, this was kind of one of the breakthroughs that helped my subfield become popular because people are like, wow, you can play Atari.

Read the Abdul-Samad Olagunju 24:06 section.Abdul-Samad Olagunju 24:06#

Can you go into vivid detail? Not too much, because, you know, obviously, most of us are beginners, but just a little bit of detail about how it works.

Read the Matthew Taylor 24:13 section.Matthew Taylor 24:13#

Read the Abdul-Samad Olagunju 25:05 section.Abdul-Samad Olagunju 25:05#

And for any newbies who want to get into AI, specifically, what things do you think they should study, you think they should just play around with code, or they should start learning more mathematics? And I don’t know what else, maybe statistics to really learn how the algorithms work?

Read the Matthew Taylor 25:23 section.Matthew Taylor 25:23#

So I would really say it depends on what that person is excited about. So, studying stats is absolutely useful. Hacking around in TensorFlow is also useful.

But if you really like programming, you probably don’t want to start hitting your head against a math textbook, right? You want to go and do something that you have fun with, that you find, you know, rewarding.

And then if you realise, oh, I really need to get into more of the details. Or oh, this math is fun. But now I want to see it work on the computer, you know, play to your strengths first. And then if you’re if you’re enjoying it, if you’re satisfied, then you can go deeper in many different directions.

Read the Abdul-Samad Olagunju 26:08 section.Abdul-Samad Olagunju 26:08#

And what inspired your love for computer science?

Read the Matthew Taylor 26:14 section.Matthew Taylor 26:14#

I played a lot of video games when I was a kid. The more serious answer, I guess, is that I see computing science as a way of creating. So I can come up with new things and new solutions that other people hadn’t before.

And that was really appealing. So just like you think of like, kids play with Legos. Like here, I found a way of creating a virtual space that just really appealed to me.

Read the Abdul-Samad Olagunju 26:44 section.Abdul-Samad Olagunju 26:44#

Okay. And what were some of the challenges as you went through high school and your undergrad? Were there any times where you’re like, I can’t do this?

Read the Matthew Taylor 26:54 section.Matthew Taylor 26:54#

Yeah, because you’re in, I mean, well, in any in any field, you’re going to find people who are much better than you or much smarter than you. And for me, it was always reminding myself, I don’t need to be the best. And there’s going to be people who are better coders, and who are better at math than I am.

But I can still contribute. And understanding that even if you’re not the best, there’s always going to be someone who’s a little bit better than you. But if you can figure out where you can contribute, what you like doing and where, what you can do that maybe other people can’t, or haven’t yet, that’s really satisfying.

Read the Abdul-Samad Olagunju 27:32 section.Abdul-Samad Olagunju 27:32#

So find your own environment, and work towards making it the best that you can make it.

Read the Matthew Taylor 27:38 section.Matthew Taylor 27:38#

You know, if you’re not going to win—most of us are not going to win the Nobel Prize. And that’s totally fine. Because you can still help the world be a better place, you can still help other people learn more, even if you’re not the absolute best. And it took it took me a long time to recognise that.

Read the Abdul-Samad Olagunju 27:57 section.Abdul-Samad Olagunju 27:57#

Okay, so I’m now going to ask you a different question. What is the worst project that you’ve been a part of?

Read the Matthew Taylor 28:06 section.Matthew Taylor 28:06#

So I’m really down on myself. But then luckily, we found another project, which was better aligned. And then things went well for the rest of grad school.

Read the Abdul-Samad Olagunju 29:18 section.Abdul-Samad Olagunju 29:18#

So in those moments, we all face challenges, but in those moments where it’s really tough, what what is it that keeps you just pushing forward? And persevering?

Read the Matthew Taylor 29:29 section.Matthew Taylor 29:29#

At that point, it was I really want to make this work. I think I could be good at this. I think I could really enjoy it. Let me at least stick it out for a year. So right, you set some time limit. And if it’s still not working, then maybe you made a mistake and you need to pivot and try something else. But if you give up at the first thing that goes wrong, then you’re not going to succeed.

Read the Abdul-Samad Olagunju 29:56 section.Abdul-Samad Olagunju 29:56#

And what would be your advice to an undergrad? About the path that they should take, or the kind of thinking they should have is as they try and make it in this field?

Read the Matthew Taylor 30:07 section.Matthew Taylor 30:07#

Yeah, well, I guess it’s a big thing is trying to decide what field you want to be in. So what are you good at? What do you enjoy? What can you actually get paid to do?

But then don’t feel like whatever you’re choosing is necessarily closing the door. You know, because I wanted to be a physicist, then I wanted to be a computer programmer, then I wanted to be an AI person. I went into academia, then I went into industry, worked for a bank, then came back to academia, you know, once you make a decision, it’s not like you can never change your mind.

Read the Abdul-Samad Olagunju 30:42 section.Abdul-Samad Olagunju 30:42#

And this is just me asking, how was it like working in a bank?

Read the Matthew Taylor 31:04 section.Matthew Taylor 31:04#

So one thing is that there’s a culture difference between like the normal bank where they were like suits. Yeah, and then there are nerds who wear a t shirt and some shorts. But there are a lot of a lot of people thinking far ahead. And it’s interesting, because the banks are usually the more conservative bunch, but then there’s also like hedge funds, who go and are much riskier and much forward thinking.

But the banks do realise that if they, if they let Amazon and Apple provide all their services, and all the good investing happens at the hedge funds, then then they’ll cease to exist. So they, you know, these are institutions that have been around for hundreds of years, and are aware that they have to change. Because you know, things like blockchain, are going to eat their lunch if they don’t provide useful services.

Read the Abdul-Samad Olagunju 31:56 section.Abdul-Samad Olagunju 31:56#

And what do you believe about the future of blockchain?

Read the Matthew Taylor 32:04 section.Matthew Taylor 32:04#

I don’t know. No, no, no, it’s, uh, I just, I know, when Bitcoin was coming out, I said, this is a scam. And man, I should have invested. So now I doubt—I just don’t trust myself.

Read the Abdul-Samad Olagunju 32:17 section.Abdul-Samad Olagunju 32:17#

Even I, a year or two ago, I remember when I was in high school, Bitcoin went up to like, 20,000. I was telling all my friends that this is a scam. And then a couple months ago it was like 70,000. Yeah. It’s crazy.

But yeah, I think I think technology is a lot more fun than people understand. I think it’s just a nerdy thing, or you can’t get into it. But there’s a lot of interesting stuff going on there. And yeah, the ability to create stuff, I think that’s the biggest thing that draws me into it now, is the fact that you can just have any idea and you can put it up there and anyone in the world can see it.

Read the Matthew Taylor 32:59 section.Matthew Taylor 32:59#

Yeah. The barrier of entry is low. So especially with apps, yeah. Right, the barrier to getting your idea into the hands of thousands of people is actually much, much lower than it used to be.

Read the Abdul-Samad Olagunju 33:11 section.Abdul-Samad Olagunju 33:11#

Exactly. So yeah, I think AI, technology, all of these things, there’s something to be excited about. But also, I think, in my opinion, I would love if there was an easier way to start learning about it.

For me, everyone talks about data analysis as a scientist, so I thought I’d take a look at it. But as you try to learn more about machine learning, data analysis, to really understand the algorithms—sometimes it feels like you’re staring at a wall, it takes a long time to start to understand the concepts. So I believe there should be an easier way to start learning it.

Read the Matthew Taylor 33:59 section.Matthew Taylor 33:59#

Well, so it absolutely can be complicated. But part of the differences, one of the questions is, are you trying to learn about it on your own? Because there’s plenty of free or cheap resources? Or are you doing it as part of some certificate or curriculum?

So right now the University of Alberta does not have a major in data science. And that’s a huge problem. So right now, we’re trying to get a major in data science put together, because we want to be able to help students quickly get into this field that has tons and tons of opportunities after you graduate.

Read the Abdul-Samad Olagunju 34:36 section.Abdul-Samad Olagunju 34:36#

Yeah. So how difficult is it to establish that?

Read the Matthew Taylor 34:41 section.Matthew Taylor 34:41#

It’s, well, it’s a lot of bureaucracy. But part of it is, I mean, between the Math Department, stats department, and computing science, we have most of the courses already. It’s more a matter of helping students find the right set of courses. So that staring at those algorithms is a little bit easier to get into.

Read the Abdul-Samad Olagunju 35:00 section.Abdul-Samad Olagunju 35:00#

So, in your opinion, working with people is always the best way to learn?

Read the Matthew Taylor 35:05 section.Matthew Taylor 35:05#

If I’m staring at an algorithm by myself, it’s just so much easier if I have someone else I can bounce ideas off of. And someone else who can tell me when I’m doing something stupid.

Read the Abdul-Samad Olagunju 35:16 section.Abdul-Samad Olagunju 35:16#

Yeah, that’s fantastic. I loved talking with you today about your history, what you believe about the future and just the way you approach problems.

And I think that a lot of people can use this kind of advice. A lot of kids, I feel like, especially me coming out of high school, you really don’t know what you want to do. But if you have an idea about like, for example, if I’m a big NBA fan, I can watch what basketball players do. Or if I’m a big football fan, I can watch the NFL and I can see what their players are doing. But it’s hard sometimes in science, going on nothing really. Making it more open would be fantastic.

Read the Matthew Taylor 35:59 section.Matthew Taylor 35:59#

Yeah. Well, and you seem to have a natural knack for interviewing people. I could see you being one of the people who helps make those bridges and tries to get nerds to be able to talk to normal people.

Read the Abdul-Samad Olagunju 36:15 section.Abdul-Samad Olagunju 36:15#

Yeah, because I think when you can explain something in an easy way to anybody then that’s when you know that you understand it. It was fantastic talking with you today. I just want to say a massive thank you.