Supervised Online Diarization With Sample Mean Loss For Multi-domain Data

Abstract

Recently, a fully supervised speaker diarization approach was proposed (UIS-RNN) which models speakers using multiple instances of a parameter-sharing recurrent neural network. In this paper we propose qualitative modifications to the model that significantly improve the learning efficiency and the overall diarization performance. In particular, we introduce a novel loss function, we called Sample Mean Loss and we present a better modelling of the speaker turn behaviour, by devising an analytical expression to compute the probability of a new speaker joining the conversation. In addition, we demonstrate that our model can be trained on fixed-length speech segments, removing the need for speaker change information in inference. Using x-vectors as input features, we evaluate our proposed approach on the multi-domain dataset employed in the DIHARD II challenge: our online method improves with respect to the original UIS-RNN and achieves similar performance to an offline agglomerative clus

Stats

• citations20
• S2 citations—
• github stars0
• HF likes0
• heat score9.92
• arxiv keyfini2019supervised

Related papers

• Fully Supervised Speaker Diarization (2018)15.80
• A Reinforcement Learning Framework For Online Speaker Diarization (2023)0.00
• Speaker Diarization As A Fully Online Learning Problem In Minivox (2020)0.00
• Speaker Diarization Using Deep Recurrent Convolutional Neural Networks For Speaker Embeddings (2017)9.41
• Enhancements For Audio-only Diarization Systems (2019)0.00
• Overlap-aware Low-latency Online Speaker Diarization Based On End-to-end Local Segmentation (2021)10.35
• Improved Large-margin Softmax Loss For Speaker Diarisation (2019)6.34
• Semi-supervised Multi-channel Speaker Diarization With Cross-channel Attention (2023)2.26