Hack Harassment: a New Initiative to Enable Communities to Reduce Online Harassment

*
Proposal
Short Form
Beginner

Excerpt

In this presentation, we will present the methodology used to create a harassment dataset and classifier, the dataset used to help the system learn what harassment looks like, along with a call to action for anyone interested to get involved with the project directly.

Description

This presentation introduces Hack Harassment, a new initiative to use machine learning to identify online harassment in social media and comment streams. Currently, online harassment goes extremely under-reported due to strict reliance on humans to report harassment and fill out a form providing context. In addition, the time for moderators to respond and apply human judgment can take days when response times in terms of minutes are needed. Some of the major social media companies have been doing ample work in automating the detection of harassment, but there is nothing available to the public, much less anything that generalizes across platforms.
 
This project provides a scalable solution for identifying online harassment in a shared, collaborative context. In this presentation, we will present the methodology used to create a harassment dataset and classifier, the dataset used to help the system learn what harassment looks like, along with a call to action for anyone interested to get involved with the project directly.

Tags

machine learning, online harassment

Speaking experience

Answered Q&A for Twitch live stream https://www.twitch.tv/videos/117579533, Gamestorm 2017 panelist, have presented to executives at tier one telecom companies, this talk will focus on technical solutions and hasn't been given before.

Speaker