Yiyuan Yang

About

This image has an empty alt attribute; its file name is 0CE1316E-E0B6-4A29-B246-90BA2005BF5C_1_105_c-1-683x1024.jpeg

I spent my undergraduate life at Vanderbilt University (2016-2019), and I majored in Mathematics and Computer Science.

I’ve always believed in the mutual influence between technology and ideology. New ideas, values, and worldviews emancipate people from the shackles of mind and allow us to envision everything in new ways and thus permit technological advances to be deployed. In turn, newer technologies radically transform societies and thus rip older institutions and barriers apart and destabilize the forces that shackle the minds.

Currently, I am an ML Generalist working on Facebook ads. My projects try to model users’ Facebook activity intents.

Academically, I’ve spent time in Medical Imaging, Computer Vision, 3D vision, multi-modal/multi-tasking learning. I had terrific experiences collaborating with Professor Landman of the MASI lab and Professor Kunda Maithilee of AIVAS Lab.

My take on Machine Learning and AI: Asking machines to learn like humans is not to teach machines to discern “truth” from “falsehood”. There is no objective truth in this world. The world, in its most natural form, is not supposed to be categorized, segmented, or understood. It’s simply a jumbo mess of basic molecules following certain natural laws that we can never fully re-express. Therefore, categorizations, segmentations, and rationalization are the ways humans “overfit” this world. Therefore, to me, artificial intelligence means “to teach machines to overfit like humans”, and I don think we can have autonomous agents overfit the world like us if we give them only rationality but not irrationality, the brain but not the body, the joy but not the pain, the jubilation but not the regret and etc.

In my spare time, I love watching talks, reading history, philosophy related books, or classical novels. My favorite sports are tennis and swimming, but I usually just go for a run for exercise.

My Favourite Books: My favorite novel right now is One Hundred Years of Solitude from Gabriel García Márquez. My favorite philosophy book is Plato’s The Republic. The history book (series) of choice is Eric Hobsbawm’s trilogy on the “long nineteenth century” (The Age of Revolution, The Age of Capital, The Age of Empire)

Photography: Be sure to check out my tiny photographic gallery as well! I usually go on a photographic trip every year. I have not decided where I should go in 2020 yet and the pandemic is kind of in the way.

Contact

Email: yiyuan.yang at vanderbilt.edu

Resume: https://drive.google.com/file/d/12Ppg76O0NNI9_TX-PfS9AOY18sOTFViM/view?usp=sharing

Linkedin: https://www.linkedin.com/in/yiyuanyang/

github: https://github.com/yiyuanyang

Research Papers

2020

Line Tracing Attention: Does time constraint force attention reprioritization? Yiyuan Yang*, Kenneth Li*, Fernanda Eliott, Maithilee Kunda. 2020 (In preparation)

VU-ObjectTrace: Time Constrained Image Tracings. Kenneth Li*, Yiyuan Yang*, Fernanda Eliott, Maithilee Kunda. 2020 (In preparation)

Validation and Estimation of Spleen Volume Via Computer-assisted Segmentation on Clinically Acquired CT Scans. Yiyuan Yang, Yucheng Tang, Riqiang Gao, Shunxing Bao, Yuankai Huo, Matthew T. McKenna, Michael R. Savona, Richard G. Abramson and Bennett A. Landman. 2020 (Under journal submission)

2019

Internal-transfer weighting of multi-task learning for lung cancer detection. Yiyuan Yang, Riqiang Gao, Yucheng Tang, Sanja L. Antic, Steve Deppen, Yuankai Huo, Kim L. Sandler, Pierre P. Massion, Bennett A. Landman. SPIE 2020. (Accepted, received honorary mention best student poster award).

* denotes equal contribution.

Computer Vision Projects

Assignments From UT Austin: CS376 Computer Vision

Analyzing Seam Carving Algorithm

Using Color, Texture and Hough Transform for Segmentation and Detection

Assignments From MIT: Advances In Computer Vision

Depth Sensing Using Projector and Camera

Hybrid Images and Motion Magnification

CS231n from Stanford: Convolutional Neural Networks for Visual Recognition

Private Github: https://github.com/yiyuanyang/stanford_cs231n

My own stuff

I tried out evolution as a learning policy early 2020: https://github.com/yiyuanyang/evolution, but doesn’t seem to improve from just using good old adam.

Some Cool Resources

We are living in an age of information explosion and unequal access to knowledge has been drastically reduced by the invention of internet (And of course, Youtube!). This means all of us have the opportunity to share and obtain knowledge whenever and wherever we want. There are some really smart and cool people out there sharing their insights and perspectives and we can stand on their shoulders to make our own decisions and form our own opinions.

That said, here are some resources I found really useful.

CS/Math Related:

  • CS231n Convolutional Neural Networks for Visual Recognition from Stanford. It gives you the opportunity to nail down all the basics of neural networks.
  • Missing Semester from MIT. It teaches Bash, Vim, Git, and all the fun stuff. If you aren’t familiar with some of them, this course can greatly improve your productivity.
  • Gilbert Strang lectures on Linear Algebra from MIT. Probably the most intuitive linear algebra course I’ve found so far and Professor Strang sounds like such a kind and nice person. I had to figure those understanding myself during freshman year because while Ph.D. students are very familiar with the content, they aren’t the best teachers on those subjects.
  • Matrix Methods in Data Analysis, Signal Processing, and Machine Learning from MIT. Taught also by Gilbert Strang.
  • 3Blue1Brown on various cool math stuff. It has THE BEST visual intuitions on many of the often neglected mathematical concepts, like Eigenvalue/vectors, Fourier transformations and etc. This makes me feel jealous, I’m sure students in the future have far better “textbooks”.
  • Computerphile on a broad range of CS concepts. Fun videos to watch when you have nothing to do.
  • TwoMinutePapers on lots of computer graphics and machine learning papers. While it does not explain all the technical stuff in detail, it does cover some of the important advancements in the aforementioned field.

Books and Reads:

  • Deep Learning from Ian Goodfellow, Yoshua Bengio and Aaron Courville
  • Artificial Intelligence from Stuart Russell and Peter Norvig

Other cool channels:

  • CrashCourse. This is probably one of my favorites since high school. It has courses on history, CS, economics, philosophy, psychology, biology and etc. It has some really cool animations and very intuitive explanations. Great for expanding the horizon of knowledge.
  • Scishow – Psych. Tons of videos on Psychology. It helps us understand how our own minds work.