Google Bindi Summit 2023
and where to
Phd Student @IIIT-Delhi
Subsequent content may contain extreme language (verbatim from social media), which does not
reﬂect the opinions of myself or my collaborators. Reader’s discretion is advised.
Workﬂow for Analysing and Mitigating Online Hate Speech
: Tanmoy and Sarah, Nipping in the bud: detection, diﬀusion and mitigation of hate speech on social media, ACM SIGWEB Winter, Invited Publication
Fig 1: The various input signals
(red), models (green) and user
groups (blue) involved in
analysing hate speech. 
Why Context is important for Hate Speech?
● No clear deﬁnition of hate speech.
● Subjective annotation from the point
of view of NLP modeling.
● No standard list of vulnerable groups.
Exogenous Signals: News articles & Topical Aﬃnity
Fig 1: Hatefulness of diﬀerent users towards diﬀerent
hashtags/topics in RETINA  Fig 2: Exogenous attention Model RETINA 
XN: News Headline
XT: Incoming Tweet
: Masud et al., Hate is the New Infodemic: A Topic-aware Modeling of Hate Speech Diﬀusion on Twitter, ICDE 2021
Fig 1: Motivation for Auxiliary Data Signals in Hate Speech Detection 
Endogenous Signals: User’s interaction on platform
: Kulkarni et al., Revisiting Hate Speech Benchmarks: From Data Curation to System Deployment, KDD 2023
More about our work:
_themessier on Twitter
lcs2lab on Twitter