Abstract

Since the proliferation of social media usage, hate speech has become a major crisis. On the one hand, hateful content creates an unsafe environment for certain members of our society. On the other hand, in-person moderation of hate speech causes distress to content moderators. Additionally, it is not just the presence of hate speech in isolation but its ability to dissipate quickly, where early detection and intervention can be most effective. However, for online platforms to become genuinely safe and democratic, it is equally important to understand what causes the generation of hate speech in the first place. Can psycho-linguistic analysis of users guide us better? This multifaceted nature of hate speech has already piqued the interest of the data mining and machine learning communities. Through this tutorial, we will provide a holistic view of what the research community has explored so far, the existing success and limitations of these approaches, and what the future holds for combating online hate speech.

Tutorial Outline

  1. Hate Speech (20 mins)
    • Definitions and motivations
    • Resources: Datasets and annotation guidelines
    • Challenges
  2. Hate Speech Detection (45 mins)
    • Modality
      • Unimodal
      • Multimodal
    • Types of the frameworks
    • Limitations
  3. Hate Speech Diffusion (40 mins)
    • Cascade prediction
    • Prediction of spreaders and consumers of hate speech
    • Role of user and network metadata
  4. ---- Break (30 mins) -----
  5. Psychological Analysis of Online Hate Spreaders (30 mins)
    • Role of user and network metadata Personality models
    • Values models
    • Empathy model
    • Conformity bias
  6. Intervention Strategy (30 mins)
    • Early detection
    • Counter hate speech
    • Proactive and reactive strategies
  7. Bias in Hate Speech Analysis (25 mins)
    • Data bias
    • Model bias
    • Other types of bias
  8. Future Scopes (20 mins)
    • Fine-grained hate speech classification
    • Sophisticated models - zero-shot/few-shot learning
    • Language-agnostic and culture-agnostic modeling
    • Fairness and bias

Presenters

Avatar

Sarah Masud

Email: sarahm [at] iiitd.ac.in

Homepage: https://sara-02.github.io

Avatar

Pinkesh Badjatiya

Email: pbadjati [at] adobe.com

Homepage: https://pinkeshbadjatiya.github.io/

Avatar

Dr. Amitava Das

Email: amitava.das2 [at] wipro.com

Homepage: http://www.amitavadas.com

Avatar

Dr. Manish gupta

Email: gmanish [at] microsoft.com

Homepage: http://research.microsoft.com/en-us/people/gmanish/

Avatar

Dr. Vasudeva Varma

Email: vv [at] iiit.ac.in

Homepage: https://faculty.iiit.ac.in/~vv

Avatar

Dr. Tanmoy Chakraborty

Email: tanmoy [at] iiitd.ac.in

Homepage: http://faculty.iiitd.ac.in/~tanmoy

Lab Page: http://lcs2.iiitd.edu.in/

Tutorial Material

Coming Soon

Contact

Email sarahm [at] iiitd.ac.in for any query about the tutorial.