Search papers, labs, and topics across Lattice.
3
0
3
4
FAD-SA-GRU achieves a remarkable 93.2% accuracy in detecting hate speech in the complex Algerian dialect, outperforming all existing models.
Achieving high performance in speech-to-speech interaction for a low-resource dialect, Dziri Voicebot sets a new standard for conversational systems in underrepresented languages.
A hybrid rumour detection framework outperforms larger models by leveraging domain-specific pre-training, achieving an impressive F1-score of 0.84 in low-resource Algerian dialect settings.