Fuzzy C-Means - 11/22/19 - 12/13/19

These past few weeks have been slow. Dr. Hassibi could not come to the last meeting and we had a school break. In the meantime, we started to learn a new algorithm that did not have to do with graph clustering. I studied Fuzzy C-Means which was fairly easy to understand. Fuzzy c-means (FCM) is a method of clustering that allows points to be in more than one cluster. This method was developed by Dunn in 1973 and improved by Bezdek in 1981. The problem of classifying outliers gave rise to non-hard algorithms that were flexible enough to effectively deal with these points, hence the name “soft clustering”. It is also known as “soft k-means” due to its design centered around k-means. A big difference between the two algorithms is that FCM assigns a membership of every node to each cluster, enabling points to be in two or more clusters. Below is an image of FCM being used to do image partitioning. After is my video explaining what FCM is in further detail.


https://youtu.be/zr50h_91gOw

       Our last meeting at Caltech before the upcoming winter break, we met with Navid Azizan Ruhi, a graduate student. At the meeting, nothing substantial was spoken about other than what we were learning. Towards the end of the meeting, we talked about what direction we wanted to take for the future. We talked about ML applied to photos, videos, and Natural Language Processing. All of them sound promising but I am particularly interested in ML applied to bioinformatics. Whatever we find ourselves doing, I hope it is super educational and helpful in the future. Below is a video showcasing how AI is used to push studies in genomics further.

https://www.youtube.com/watch?v=uC3SfnbCXmw

       On another note, during my school break a few weeks ago, I went to San Francisco with my friend to participate in FrontierHacks. It was a 24-hour event where people would make a project and showcase it to judges. We went through a lot of troubles but the main takeaway from this experience was the experience itself. It was my first time organizing flights, planning the trip, and traveling without my parents. It definitely was an eye-opening moment into adulthood. I am super grateful to my parents for giving me this opportunity. Without them, I couldn't have even gone there which is a privilege itself. We made a Mandarin Character classifier to approach the hackathon's education vertical. While it did not work the way it was intended to nor won any awards, it was a good experience. Below is a picture of FrontierHack's venue.

Image result for make school

       Winter break, which is right after finals, will be a pretty lazy one for me. My brother is flying to Los Angeles from New York to visit us which is great. I plan on doing some side projects and get a head start in some of my classes.

Comments

Post a Comment

Popular posts from this blog

Technical Journal - Research Papers

Image
On the Efficiency of Data Collection for Crowdsourced Classification :        The paper aims to explain the accuracy gaps between popular collection policies (state-action mapping) on crowdsourced data. The quality of crowdsourced data is often highly variable so it is suspected that it is the cause. However, studies show that the policies used to collect such data have a strong impact on the accuracy of the system. The first theoretical explanation of the accuracy gaps is that the non-adaptive uniform allocation, and the adaptive uncertainty sampling and information gain maximization. Done with the representation of the collection process in terms of random walks in the log-odds (log of odds of guessing correctly/incorrectly) domain, the author derives lower and upper bounds on the accuracy of the policies. The bounds accurately tell us by how much do two adaptive policies trump a non-adaptive one. It is believed that techniques used by them are applicable to addi...
Read more

Journal

Image
On Tuesday, March 24th, we had a meeting with Professor Hassibi over Zoom. Prior to our talk, we had issues with our code regarding the fact that it didn't produce the results expected. He told us to find and graph the eigenvalues of the graph Laplacian, as that would indicate the number of clusters found and value you can refer other datapoints to. He noted that graphing the corresponding eigenvector would indicate whether a cluster had been isolated or not. On that note, it was up to us to tinker with values for a bit and try different algorithms like K-means. Below is a video of the meeting with Professor Hassibi: This week has been very interesting, to say the least. As you may know, COVID-19 has just taken the U.S. for a turn and has put a pause on daily life, forcing businesses to adapt, overcome, or fail in some cases. For us high schoolers, classes have migrated to the Zoom platform. With it followed teachers with bad internet service, leading to indecipherable dialect...
Read more

First Month

Image
The first four weeks of this Caltech Stem class have been interesting, to say the least. Much of the material we went into have been all over the place with various objectives. We started off with teamwork exercises, watching college students present their research, replicating one for ourselves, writing a concept map, and going to Caltech to meet Dr. Hassibi. We had two, teamwork workshops for the first two weeks. One of which was the dominoes problem where each student was given two dominoes and we were instructed to get three dominoes per person by exchanging dominoes for chips and chips for dominoes. Instructions were very vague other than that everyone must have a domino to "survive" a round. We struggled in the first few attempts but through several days of discussion and trial & error, we were able to find a way to accomplish our goal. Some key underlying goals of this exercise were the importance of leadership, communication, and open-mindedness...
Read more