# Improving Speaker Identification for Shared Devices by Adapting Embeddings to Speaker Subsets

arXiv: https://arxiv.org/abs/2109.02576 · HTML: https://mlsanity.com/paper/2109.02576/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 6 Sep 2021.

## Built on by (most cited first, 3 of 3)

- [Parameter-Free Attentive Scoring for Speaker Verification](https://mlsanity.com/paper/2203.05642/) (2022)
- [Graph-based Multi-View Fusion and Local Adaptation: Mitigating Within-Household Confusability for Speaker Identification](https://mlsanity.com/paper/2207.04081/) (2022)
- [Probabilistic Back-ends for Online Speaker Recognition and Clustering](https://mlsanity.com/paper/2302.09523/) (2023)

## Abstract (arXiv)

Speaker identification typically involves three stages. First, a front-end speaker embedding model is trained to embed utterance and speaker profiles. Second, a scoring function is applied between a runtime utterance and each speaker profile. Finally, the speaker is identified using nearest neighbor according to the scoring metric. To better distinguish speakers sharing a device within the same household, we propose a household-adapted nonlinear mapping to a low dimensional space to complement the global scoring metric. The combined scoring function is optimized on labeled or pseudo-labeled speaker utterances. With input dropout, the proposed scoring model reduces EER by 45-71% in simulated households with 2 to 7 hard-to-discriminate speakers per household. On real-world internal data, the EER reduction is 49.2%. From t-SNE visualization, we also show that clusters formed by household-adapted speaker embeddings are more compact and uniformly distributed, compared to clusters formed by global embeddings before adaptation.