Twitter does not want to put an edit button in order to prevent the abuse. What could go wrong? Can we learn this through analysis of retrospective edits on other platforms? Facebook allows edits and keeps log of the old versions of posts. Youtube allows retrospective edits but does not keep such a log.
The project will have the following steps:
1- Collect a dataset of posts and comments from Youtube and other platforms
2- (If possible) summarize what’s changed in the original content.
3- Describe changes with topic modeling or another clustering method
4- Annotate and classify malicious edits (e.g. spam)
Bachelors are welcome to pick this project.
You can also work as a group of two.
Requirements: Basic ML knowledge and data analysis skills in Python.
Please send a C.V. and Transcript!
You do not need to submit anything else beforehand.
| Contact: | Tuğrulcan Elmas |