Final's Week (12/2 and on)
December 8
- finished research proposal paper!
- Graduate student interview (interviewed Animesh!)
- How did you decided to go to graduate school? How did you decide on a masters'/PhD over the other option?
- He wanted to be professor
- Why be a professor?
- Because it's the only job where you can both teach and research.
- Did you work before going to graduate school?
- Do you ever consider working in industry?
- How did you select UCSB? What factors were important to you?
- He looked for research that was happing in distributed systems and edge computing, UCSB has Racelab which is cool!
- What undergraduate experience/course work/project work was most important in helping prepare you for graduate work and research?
- He had undergrad internships - there was a student run organization called microsoft innovation labs through his college, which exposed him to research and academia.
- Which undergrad did you attend?
- Is graduate school what you thought it would be like? What is it like? What did you expect? How are these the same/different?
- Graduate school is exactly what he thought it would be like (it is very rough).
- Tell me about a time that you've been frustrated or discouraged with your research? How did you get through it? What did you do?
- Once he was frustrated because the research wasn't what he wanted to do research on. Plus the professor was not flexible to let him do what he wanted. The solution: build the research into what he wanted to do.
- What's your favorite part about doing research? Tell me about a good research experience you have had.
- Favorite part: when the things start to click and everything fits in and the results start rolling in, whether it be positive or negative. Good experience: He was working on two separate projects and was in a weird position, so he didn't have any control over the second project. But, what ended up happening is that towards the end of his contract, the other lead developer suddenly left so the professor said to continue his work that he left. Turns out the guy that left didn't do any of the work, so Animesh discovered that the features weren't working. Animesh had an opportunity to improve the project and one day everything clicked and fell into place and the project was working!
- What's your favorite aspect of graduate school (the whole thing, not just research or classes)? Your least favorite?
- Most favorite: theres encouragement to pursue hobbies outside of research in labs. Least favorite: when things don't work out within the time that you have. Let's say you have a paper deadline in a month, everyone does their best to finish it. but in the end you run out of time and it is unable to be published.
- What's your relationship like with your adviser?
- Not much more than a professional relationship because he's only been in the lab for one quarter.
- How early were you exposed to CS?
- Exposed during high/middle school. He had an extremely bad programming class but that was what sparked his curiosity. He got really interested during college, though.
-
Interview reflection:
Week 10 (11/25-12/1)
Weekly goals:
- [ ] continue labeling camera trap images
- [x] attend cs 110 lecture
- [x] finish research proposal
- [x] create and practice presentation
- [x] continue researching state of the art methods
December 1 (3.5 hours)
- cs 110 lecture
- give presentations to other groups
- fill out feedback form for other groups presentations
- undergrad meeting
- continue editing research proposal
- grad mentor group meeting
- mentors did not end up showing up :(
November 30 (2 hours)
- lab group meeting (ML presentation by Animesh)
- three main models we can focus for wtb: classification (identifying entire image), detection (having bounding boxes), segmentation (knowing which part of the image belongs to which class, resulting in tight, free form bounding box over recognizable aspects of the image)
- diff models for image classification: MobileNet, Inception, ResNet
November 29 (3 hours)
- cs 110 lecture
- receive feedback from ERSP instructors for our proposals
- include more background information
- go into detail about what specific machine learning algorithms we're using
- our images are nice!