Specialization Spotlight: Machine Learning
What’s your year in school, specialization, and position?
Samruddhi Hande: I’m a third-year specializing in Machine Learning. I’m also the Co-President of CSSA.
Phillip Lagoc: I'm a fourth-year specializing in Machine Learning and I’m the internal public relations chair for CSSA.
Mehail Sunny Mathew: I am a third-year specializing in Machine Learning and I am also the Sponsorship Chair.
What’s a fun fact about you?
SH: I’ve lived in San Diego my entire life.
PL: My mom owns a dragon fruit farm in the Philippines, and that farm is actually, according to her, it’s pretty well known. She like went to the Philippines and she was on TV for a bit about that farm.
MSM: I played golf for four years in high school.
What got you into Cognitive Science and your specialization?
SH: I actually applied to colleges as a computer science major and I got into some other schools for computer science but when I came here and I heard about Cognitive Science at UCSD, it was really interesting to me because I wouldn’t be just doing computer science classes, I’d be taking psychology, computer science, linguistics — all these different classes. Cognitive Science is a very interdisciplinary major and I like that aspect of it. I chose Machine Learning because I was interested in the technology side of Cognitive Science. I think that’s fit really well for me because I get to combine Neuroscience and Psychology with the Computer Science aspect of Cognitive Science.
PL: I remember I went to CSSA’s first GBM as a sophomore and I just saw that it was really interesting. That made me realize what CogSci was and then when I took Cogs 1, I realized I really liked it. As for my specialization: my roommate was really into machine learning and told me more about it.
MSM: My dad was telling me about what majors were going to be useful in the future and one of them was Cognitive Science. I started to look into it more and picked machine learning because I was always interested in topics such as artificial intelligence.
What’s the most important thing you have learned in this field?
SH: I guess the Data Science part of Machine Learning. For me, that’s been the most useful because I see it everywhere and it’s the career path that I want to follow and it’s very relevant to the job market.
PL: I think the most important thing I’ve learned was how to program in Python, and just understanding the math behind machine learning like linear algebra, matrix multiplication, vector calculus and all that. Another thing I learned is just to talk to your professors; ask them for help because they’re really willing to help out.
MSM: To be honest, I haven’t taken many machine learning classes yet, because a lot of them have prerequisites so I’ve been working on that.
What COGS classes did you find the most useful?
SH: As I said I found COGS 108 and COGS 109 the most useful so COGS 108 is “Data Science in Practice” and COGS 109 is “Modeling and Data Analysis”. In both of those, we discussed techniques of data analysis like learning how to build certain models, learning what those models mean and what insights you can grab from them. They were also very hands-on programming classes (when we used Jupyter notebooks) and it’s always better to learn by doing.
PL: The COGS 118 series mainly because they’re just a basic introduction to machine learning. As for the math, I would say MATH 20E, MATH 180A, and COGS 109, because it gives you a better introduction to all of the statistical methods you need in machine learning. I would also recommend taking online classes like the Coursera course about machine learning by Stanford Professor, Andrew Ng.
MSM: I feel like COGS 17 was kind of useful because you learn a lot about the neuroscience aspect. And I kind of had like a foundation of like how the mind works in general. I feel like it’s good to know for all specializations because like cognitive science is kind of centered about how the mind works and applying that to different like aspects of like life.
What’s your favorite COGS class and why?
SH: I liked COGS 108 because Professor Shannon Ellis was really nice and made the concepts easier to understand. It also introduced me more into the field of data science.
PL: COGS 109 was one of my favorite classes because it was eye-opening because I got to really dive deep into the math behind machine learning.
MSM: COGS 101A was pretty fun because we learned a lot about visual illusions and like tricks that our mind plays. We did a lot of labs that actually like show it so it was really interactive.
Who’s your favorite COGS professor?
SH: From professors that I’ve had, I’d say Professor Shannon Ellis, just because she’s very easy to talk to, she’s very youthful, and she’s really easy to approach. I was also a TA for her COGS 9 class in the Fall.
PL: I liked Professor Ellis. She taught cogs 108 and she’s just a super nice, helpful and outgoing.
MSM: I think Tayler Scott was my favorite professor because he was really relatable. He talked a lot about Reddit and memes and explained concepts really well. I took him for COGS 10, which was about the consequences of technology.
What did you do last summer?
SH: Last summer, I interned at a startup. I was on a conversational AI team, so basically my team of three made an Alexa skill to help users keep track of their habits and the entire startup is based on creating Alexa skills based on a personality. We created an avatar named Harmony, who serves as like a personal assistant to remind you to track your habits or when you’re not doing your habits. I just learned how to make an Alexa skill, which is pretty cool.
PL: I was doing research for my machine learning lab on campus and also took CSE 21 during the summer session.
MSM: I’ve been taking summer sessions the past two summers to get a lot of the prerequisites out of the way. I’ve also been taking online courses in machine learning. I’m currently applying for internships for this upcoming summer.
Are you involved in any research or other activities on campus?
SH: I did a lab the summer after my freshman year, which was a natural computation lab. I think through that I did learn a lot and I enjoyed it but it showed me that research was up for me so I stayed away from that. Then more hands-on things, but other than academics, this year I think Cognitive Science Student Association is my biggest commitment.
PL: I work at the machine learning for a social science lab. At that lab, they’re part of the Center for Peace and Security Studies, which is under the political science department. And what essentially we do is we use machine learning to answer questions that social scientists have, like what are the causes of war and peace, stuff like that. So it’s a pretty interesting intersection between, like the technical side of Science and Social Sciences. I’m also an I.A. for some classes.”
MSM: I haven’t yet but I am currently looking for labs in neuroscience or machine learning.
What advice or resources do you have for first years and people starting in the specialization or cognitive science in general?
SH: I was very conflicted about whether to go for Machine Learning because it seemed very scary to me and also it’s a lot of Math, which I did enjoy in high school, but obviously, as it gets harder, it’s hard to enjoy it. I think for the biggest piece of advice I can give, just stick with it because in the end, it’ll definitely be rewarding; everyone wants people who know Machine Learning these days because you can apply it to so many industries. Don’t be scared by what machine learning looks like and all the courses you have to take for it.
PL: I would say be comfortable with math. Also take those online courses, especially that Coursera course. Don’t be afraid to ask your professors for help, as they’re really willing to help you out.
MSM: It is important to focus more on the prerequisites and core classes for your specialization first, rather than taking GEs because you can do those anytime. Taking prerequisites and specialization courses early on would help you gain experience and be more marketable for internships.
What’s your favorite part of CSSA?
SH: I think the best part about CSSA is that I really learned more about the cognitive science program at UCSD because it’s so big because there are so many specializations. I think like being on the board and being a part of it for two years now has really shown me more opportunities and it’s definitely connected me well with people. Even my internship this past summer at a startup I got through being on the board because the startup reached out to us to advertise it. I don’t know if you know Phillip, but he’s the PR chair on board and he’s a year older than me so he’s been a valuable resource for me in terms of the classes and just someone who has taken what I’m taking right now and getting that kind of future insight.
PL: My favorite part is that it brings together a whole bunch of people with similar interests and all that into one community that’s really helpful and friendly. It’s not only like newcomers to that community, but also people who are already within it. I would just say the inclusivity of the CSSA was one of my favorite things.
MSM: All the new people I’ve met, including the executive board and students who show up at our events. I‘ve gotten to talk to people I otherwise wouldn’t have gotten the chance to.
What would you do differently if you could go back in time?
SH: I think if I go back in time, I would take certain classes earlier, because like I said my biggest advice was not to be scared and for a long time I was scared of taking MATH 180A. Now I have other hard classes to take too so I wish I’d taken it earlier so that I could have started the COGS 118 series earlier too, and just not being afraid to jump into these hard classes and then I could have timed it out a little bit better.
PL: If I could go back in time, I would plan my career out more because since machine learning is such a competitive field, you want to have like a plan and start as soon as possible. For me, I didn’t take most of these upper-division classes until I was an upperclassmen. But I would say just having a plan as early as possible, starting early and also going out there and asking professors. I was really shy back then and now I regret that.
MSM: Freshman year, I wasn’t sure I wanted to do machine learning specialization so I took MATH 10 series instead of 20. I took Math 10B and 10C before I realized so I had to go back and do MATH 20B and 20C in the community college over the summer.
What are your future plans after graduation?
SH: I’m doing a lot of software engineering right now but I want to go into product management in the long run. I like coding but I’m definitely more of a people person. I want to use that coding and programming knowledge in a strategic manner. I don’t want to just sit there and code all day. My goal is to get the software engineering background, specifically data science and machine learning applications, and then use that in product management.
PL: For me, I want to go into industry for at least a year or two, to just hone those technical skills. And then after that, I want to go to grad school, get my masters and potentially get a Ph.D.
MSM: I’ll probably try to work in the industry. I’ve been looking at both companies and startups. I might go to grad school later, but I want to get some experience in the real world. Grad school will always be there in the future.
