Federated Marginal Personalization for ASR Rescoring
This work addresses the challenge of efficiently learning personalized NNLMs on devices with privacy constraints for ASR rescoring, offering an incremental improvement over existing federated fine-tuning methods.
This paper introduces Federated Marginal Personalization (FMP), a method for continuously updating personalized neural network language models (NNLMs) on private devices using federated learning. Instead of fine-tuning NNLM parameters, FMP adjusts word probabilities using adaptation factors derived from global and personalized marginal distributions, achieving modest word error rate (WER) reductions on two speech evaluation datasets.
We introduce federated marginal personalization (FMP), a novel method for continuously updating personalized neural network language models (NNLMs) on private devices using federated learning (FL). Instead of fine-tuning the parameters of NNLMs on personal data, FMP regularly estimates global and personalized marginal distributions of words, and adjusts the probabilities from NNLMs by an adaptation factor that is specific to each word. Our presented approach can overcome the limitations of federated fine-tuning and efficiently learn personalized NNLMs on devices. We study the application of FMP on second-pass ASR rescoring tasks. Experiments on two speech evaluation datasets show modest word error rate (WER) reductions. We also demonstrate that FMP could offer reasonable privacy with only a negligible cost in speech recognition accuracy.