Scholarly-X
Unmute
Mentor: Aditi Kabra, Software Engineer (IC2) at Microsoft
Team members: Edward Haas, Char Wae (Erica) Chong, Jovan Petrovic
When trying to familiarize yourself with a new technical subject the best way is often to read seminal papers.
The question is how to identify the seminal papers.
Suppose you are given a single reference as a starting point. How do you figure out what else to read?
Quoting an earlier writeup, "Think of the literature of a field like a tree. There are root papers that inspire child papers, and the most recent papers are leaf nodes. The objective is to traverse the tree in such a way as to get a decent idea of it’s general shape, and conceptual understanding of its trunk and thickest branches so that you have the context to comprehend even the leaf nodes. You can climb both up and down a tree, so it’s a tree traversal problem. If you start close to the leaves, you go back to an ancestor, and then look at its most important children. If you start at the root, you do the opposite, and look for successor papers."
The objective of the project would be to write software to use google scholar or maybe some other research paper api to automatically using citation counts, paper relationships, and other metrics, find and identify the seminal papers of a field. The final result could be put up on github or hosted somewhere- it is something my friends and I thought would help our lives a lot. A potential expansion can be to use machine learning.
Aforementioned writeup to explain the motivation in further detail: https://aditink.github.io/literature/2020/04/12/literature-review.html
Team members: Edward Haas, Char Wae (Erica) Chong, Jovan Petrovic
When trying to familiarize yourself with a new technical subject the best way is often to read seminal papers.
The question is how to identify the seminal papers.
Suppose you are given a single reference as a starting point. How do you figure out what else to read?
Quoting an earlier writeup, "Think of the literature of a field like a tree. There are root papers that inspire child papers, and the most recent papers are leaf nodes. The objective is to traverse the tree in such a way as to get a decent idea of it’s general shape, and conceptual understanding of its trunk and thickest branches so that you have the context to comprehend even the leaf nodes. You can climb both up and down a tree, so it’s a tree traversal problem. If you start close to the leaves, you go back to an ancestor, and then look at its most important children. If you start at the root, you do the opposite, and look for successor papers."
The objective of the project would be to write software to use google scholar or maybe some other research paper api to automatically using citation counts, paper relationships, and other metrics, find and identify the seminal papers of a field. The final result could be put up on github or hosted somewhere- it is something my friends and I thought would help our lives a lot. A potential expansion can be to use machine learning.
Aforementioned writeup to explain the motivation in further detail: https://aditink.github.io/literature/2020/04/12/literature-review.html
What tools did you use to create your project?
How much experience does your group have? Does the project use anything (art, music, starter kits) you didn't create?
CodeDay Labs advanced-track team
Made with 💖 by robots running our open-source software and GraphQL APIs.