The Science Behind Detecting Hate Speech on Social Media

[anjalibhavan]

• Abstract (2-3 lines)
  The concepts, characteristics and study techniques of hate speech on social media, particularly on Twitter. Prevalent research and future scenario.

• Brief Description and Contents to be covered
  As times go, this is a particularly contentious time we live in, especially on the internet. There has been a proliferation of hate speech content on various social media websites, and studying and detecting them is an important problem at hand. Through this talk I aim to decode the concept and phenomenon of hate speech, the prevalent methods of analysis and the latest research. The focus will be on Twitter.
  Some points I hope to cover:
  1.Discussion of a few research papers on the topic
  2.The basic concepts behind hate speech analysis, what sets it apart from regular texts and how we use NLP to process and classify these texts.
  3.Aspects of hate speech analysis - demography based, targeted hate, individual-centric hate etc.
  4.A brief look into my recent research on hate speech detection

• Pre-requisites for the talk
  Machine learning basics preferred, but not much else.

• Time required for the talk
  20 minutes.

• Link to slides
  https://docs.google.com/presentation/d/1Fui1zfJ2dtgOUnHdRMCDx8d6yT2lno-b7cfMPGilrfg/edit?usp=sharing

• Will you be doing hands-on demo as well?
  No

• Link to ipython notebook (if any)
  Not applicable

• About yourself
  Incoming MS Computational Linguistics student at University of Washington. Software Engineer at PayTM and NLP researcher.

• Are you comfortable if the talk is recorded and uploaded to PyData Delhi's YouTube channel ?
  Yes

• Any query ?
  No

[shagunsodhani]

Thanks for the proposal @anjalibhavan It looks quite interesting. Could you ping me when you have uploaded the slides?

[anjalibhavan]

@shagunsodhani I've added the slides. In the presentation there's an analysis of just one paper. Let me know if I should add discussion for another one?

[shagunsodhani]

Slides look good. I would suggest blurring the name/profile of the users in twitter screenshots. Best of luck with the talk.

[anjalibhavan]

@MSanKeys963 When do you think I can present this talk?