|Education:||Indiana University Bloomington|
|Mentor:||Dr. James Abello|
The main goal for this project is developing algorithms to decompose it in order to extract semantics and translate these massive datasets into a “human understandable story” that describes it from massive datasets.
For the first week of DIMACS REU program, it was started with the orientation via Cisco. It is such a precious chance for me to have research experience as an undergraduate student and it is going to be done remotely. I had attend the first workshop of HTML to make sure I am on the right track to create this website. The whole experience had been great and I'm looking forward to meet my mentor and get started on the research!
This week started with brief presentations on the research topic from every participant in DIMACS REU program, I had present my research as well. It was nervous to me becuase due to the miscommunication casued by Gmail, I wasn't able to have time to reach my mentor, Dr. Abello, util the morning of the presenation day. However, it was an amazing experiences and I found a lot of other people' research topic pretty interesting. Also, I was able to prepare for the research and meet the team I'm going to work with.
I signed up for the data science bootcamp for this week and I had successfuly found a way to help the research by translating the 3k labels of datasets we were using from Danish to English. During the weekly meeting, we had discussed about the Parallel BFS algorithmn and the shortest path algorithmn to try to extract sematic from the dataset.
During this week, I had attend the semainer by Mykhaylo TyomkynI, it was quite interesting. I had also signed up for two Neo4j tranings for next week by Dr.Abello's recommandation. For this week's research work, I had helped my teammate to makde changes of the graph-strata system on the sparesnet display and had successfuly set up the graph-strata system on my computer so I can do research on the datas by using the system.
During this week, I had done some fixed point decomposition by using the iterative edge decomposition. Beside the research work, I had also attend the semainer by Lenka Zdeborova on "Understanding machine learning with statistical physics". Although physics is not my favorite, the semainer has casued me a lot of thinking on machine learning. In addition, I had completed the two Neo4j tranings.
This week, I had reported my research work on the findings on Datasets and Visualization to Dr.Abello by a presenation. The presentation contains findings on the similarities and differencies between the new dataset that Dr.Abello shared and the original dataset that I had been worked on. With the same amount of dataset as 1,852 stories but with 19,738 vertices in the original datasets and 52,929 vertices in the new datasets. There are more findins beside it that can be found in powerpoint listed in the references.
This week, I had been putting my effort on extract the sematics of a dataset and creating stories to describe the dataset. In addition of that, I had attend a workshop for Ethics in Research on Monday, a semainer with Vivek Singh, and a session of Graduate school panel.
I had been working on the presentation slides and final report for my REU experience during this week.
This week is the last week of the REU experience, I had present my work scheduled for Thursday and it is a great experience overall.
Abello, James & Mawhirter, Daniel & Sun, Kevin. (2019). Taming a Graph Hairball: Local Exploration in a Global Context. 10.1007/978-3-030-06222-4_10.
Abello, J., Quelroy, F.: Network decompositions into fixed points of degree peeling. Social Networks Analysis and Mining pp. 4–19 (2014).
Abello, J., & Nakhimovich, D. (2020). Graph waves. CEUR Workshop Proceedings, 2578.
Abello, J., Hohman, F., Bezzam, V., & Chau, D. H. (2019). AtlaS: Local graph exploration in a global context. 165-176. Paper presented at 24th ACM International Conference on Intelligent User Interfaces, IUI 2019, Marina del Ray, United States.
ualberta.ca. (n.d.). Unpacking World Folk-literature: Thompson’s Motif Index, ATU’s Tale Type Index, Propp’s Functions and Lévi-Strauss’s Structural Analysis for folk tales found around the world. Sites.Ualberta.Ca. Retrieved July 21, 2020, from https://sites.ualberta.ca/~urban/Projects/English/Motif_Index.htm
Wikipedia contributors. (2020, June 7). Aarne–Thompson–Uther Index. In Wikipedia, The Free Encyclopedia. Retrieved 09:20, July 21, 2020, from https://en.wikipedia.org/w/index.php?title=Aarne%E2%80%93Thompson%E2%80%93Uther_Index&oldid=961220406