Speaker-Smoothed kNN Speaker Adaptation for End-to-End ASR
Authors:
Shaojun Li,
Daimeng Wei,
Jiaxin Guo,
ZongYao Li,
Zhanglin Wu,
Zhiqiang Rao,
Yuanchang Luo,
Xianghui He,
Hao Yang
Abstract:
Despite recent improvements in End-to-End Automatic Speech Recognition (E2E ASR) systems, the performance can degrade due to vocal characteristic mismatches between training and testing data, particularly with limited target speaker adaptation data. We propose a novel speaker adaptation approach Speaker-Smoothed kNN that leverages k-Nearest Neighbors (kNN) retrieval techniques to improve model out…
▽ More
Despite recent improvements in End-to-End Automatic Speech Recognition (E2E ASR) systems, the performance can degrade due to vocal characteristic mismatches between training and testing data, particularly with limited target speaker adaptation data. We propose a novel speaker adaptation approach Speaker-Smoothed kNN that leverages k-Nearest Neighbors (kNN) retrieval techniques to improve model output by finding correctly pronounced tokens from its pre-built datastore during the decoding phase. Moreover, we utilize x-vector to dynamically adjust kNN interpolation parameters for data sparsity issue. This approach was validated using KeSpeech and MagicData corpora under in-domain and all-domain settings. Our method consistently performs comparably to fine-tuning without the associated performance degradation during speaker changes. Furthermore, in the all-domain setting, our method achieves state-of-the-art results, reducing the CER in both single speaker and multi-speaker test scenarios.
△ Less
Submitted 11 June, 2024; v1 submitted 7 June, 2024;
originally announced June 2024.
Sparse and Switching Infinite Horizon Optimal Control with Mixed-Norm Penalizations
Authors:
Dante Kalise,
Karl Kunisch,
Zhi** Rao
Abstract:
A class of infinite horizon optimal control problems involving mixed quasi-norms of $L^p$-type cost functionals for the controls is discussed. These functionals enhance sparsity and switching properties of the optimal controls. The existence of optimal controls and their structural properties are analyzed on the basis of first order optimality conditions. A dynamic programming approach is used for…
▽ More
A class of infinite horizon optimal control problems involving mixed quasi-norms of $L^p$-type cost functionals for the controls is discussed. These functionals enhance sparsity and switching properties of the optimal controls. The existence of optimal controls and their structural properties are analyzed on the basis of first order optimality conditions. A dynamic programming approach is used for numerical realization.
△ Less
Submitted 16 November, 2020; v1 submitted 31 August, 2018;
originally announced August 2018.
