CRUSH: Contextually Regularized and User anchored Self-supervised Hate speech Detection.
Parag DuttaSouvic ChakrabortySumegh RoychowdhuryAnimesh MukherjeePublished in: CoRR (2022)
Keyphrases
- speech recognition
- user satisfaction
- detection rate
- speech signal
- automatic detection
- user experience
- detection accuracy
- spoken dialogue systems
- object detection
- neural network
- user preferences
- user centric
- user interaction
- user input
- least squares
- end users
- multimodal interfaces
- website
- automatic speech recognition
- query formulation
- human communication
- speech synthesis
- voice activity detection
- audio visual
- false alarms
- user model
- user defined
- false positives
- detection method
- human computer interaction
- collaborative filtering
- relevance feedback
- hidden markov models
- face recognition
- multimedia
- information systems