# Interrupted and cascaded permutation invariant training for speech separation

arXiv: https://arxiv.org/abs/1910.12706 · HTML: https://mlsanity.com/paper/1910.12706/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 28 Oct 2019.

## Built on by (most cited first, 8 of 8)

- [Heterogeneous Target Speech Separation](https://mlsanity.com/paper/2204.03594/) (2022)
- [Optimal Condition Training for Target Source Separation](https://mlsanity.com/paper/2211.05927/) (2022)
- [On permutation invariant training for speech source separation](https://mlsanity.com/paper/2102.04945/) (2021)
- [Time-Domain Speech Extraction with Spatial Information and Multi Speaker Conditioning Mechanism](https://mlsanity.com/paper/2102.03762/) (2021)
- [Single-channel speech separation using Soft-minimum Permutation Invariant Training](https://mlsanity.com/paper/2111.08635/) (2021)
- [Progressive Learning for Stabilizing Label Selection in Speech Separation with Mapping-based Method](https://mlsanity.com/paper/2110.10593/) (2021)
- [SANE-TTS: Stable And Natural End-to-End Multilingual Text-to-Speech](https://mlsanity.com/paper/2206.12132/) (2022)
- [Improving Label Assignments Learning by Dynamic Sample Dropout Combined with Layer-wise Optimization in Speech Separation](https://mlsanity.com/paper/2311.12199/) (2023)

## Abstract (arXiv)

Permutation Invariant Training (PIT) has long been a stepping stone method for training speech separation model in handling the label ambiguity problem. With PIT selecting the minimum cost label assignments dynamically, very few studies considered the separation problem to be optimizing both the model parameters and the label assignments, but focused on searching for good model architecture and parameters. In this paper, we investigate instead for a given model architecture the various flexible label assignment strategies for training the model, rather than directly using PIT. Surprisingly, we discover a significant performance boost compared to PIT is possible if the model is trained with fixed label assignments and a good set of labels is chosen. With fixed label training cascaded between two sections of PIT, we achieved the state-of-the-art performance on WSJ0-2mix without changing the model architecture at all.