Acoustic Model Optimization over Multiple Data Sources: Merging and Valuation
Not on the map. Why?
Abstract (arXiv)
Due to the rising awareness of privacy protection and the voluminous scale of speech data, it is becoming infeasible for Automatic Speech Recognition (ASR) system developers to train the acoustic model with complete data as before. For example, the data may be owned by different curators, and it is not allowed to share with others. In this paper, we propose a novel paradigm to solve salient problems plaguing the ASR field. In the first stage, multiple acoustic models are trained based upon different subsets of the complete speech data, while in the second phase, two novel algorithms are utilized to generate a high-quality acoustic model based upon those trained on data subsets. We first propose the Genetic Merge Algorithm (GMA), which is a highly specialized algorithm for optimizing acoustic models but suffers from low efficiency. We further propose the SGD-Based Optimizational Merge Algorithm (SOMA), which effectively alleviates the efficiency bottleneck of GMA and maintains superior model accuracy. Extensive experiments on public data show that the proposed methods can significantly outperform the state-of-the-art. Furthermore, we introduce Shapley Value to estimate the contribution score of the trained models, which is useful for evaluating the effectiveness of the data and providing fair incentives to their curators.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 55 of its 112 references that are in the map. First posted on arXiv 21 Oct 2024.
Builds on
55 of 112 references.
References · 55
- 2017Attention Is All You Needcited by 60,361
- 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
- 2014Adam: A Method for Stochastic Optimizationcited by 37,240
- 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
- 2016Communication-Efficient Learning of Deep Networks from Decentralized Datacited by 7,669
- 2014Neural Machine Translation by Jointly Learning to Align and Translatecited by 4,548
- 2013Efficient Estimation of Word Representations in Vector Spacecited by 4,375
- 2014Sequence to Sequence Learning with Neural Networkscited by 2,986
- 2019ALBERT: A Lite BERT for Self-supervised Learning of Language Representationscited by 2,790
- 2018Deep contextualized word representationscited by 2,290
- 2016WaveNet: A Generative Model for Raw Audiocited by 1,722
- 2016Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translationcited by 1,475
38 more
- 2019Deep Leakage from Gradientscited by 1,021
- 2016Pixel Recurrent Neural Networkscited by 816
- 2018How To Backdoor Federated Learningcited by 765
- 2016Federated Optimization: Distributed Machine Learning for On-Device Intelligencecited by 564
- 2018Exploiting Unintended Feature Leakage in Collaborative Learningcited by 520
- 2017Deep Models Under the GAN: Information Leakage from Collaborative Deep Learningcited by 334
- 2019Towards Efficient Data Valuation Based on the Shapley Valuecited by 281
- 2018Beyond Inferring Class Representatives: User-Level Privacy Leakage From Federated Learningcited by 246
- 2016Stochastic Variance Reduction for Nonconvex Optimizationcited by 145
- 2013Hacking Smart Machines with Smarter Ones: How to Extract Meaningful Data from Machine Learning Classifierscited by 130
- 2018An Alternative View: When Does SGD Escape Local Minima?cited by 115
- 2017Unsupervised Learning of Disentangled and Interpretable Representations from Sequential Datacited by 93
- 2019Deep Learning for Audio Signal Processingcited by 92
- 2019A Survey of Optimization Methods from a Machine Learning Perspectivecited by 91
- 2019Measure Contribution of Participants in Federated Learningcited by 79
- 2019Jasper: An End-to-End Convolutional Neural Acoustic Modelcited by 76
- 2019Federated Learning for Edge Networks: Resource Optimization and Incentive Mechanismcited by 75
- 2017The Microsoft 2017 Conversational Speech Recognition Systemcited by 52
- 2020FMore: An Incentive Scheme of Multi-dimensional Auction for Federated Learning in MECcited by 48
- 2019Very Deep Self-Attention Networks for End-to-End Speech Recognitioncited by 46
- 2019Incentive Design for Efficient Federated Learning in Mobile Networks: A Contract Theory Approachcited by 45
- 2019Motivating Workers in Federated Learning: A Stackelberg Game Perspectivecited by 45
- 2019NEZHA: Neural Contextualized Representation for Chinese Language Understandingcited by 39
- 2020An Incentive Mechanism for Federated Learning in Wireless Cellular network: An Auction Approachcited by 39
- 2019Language Models with Transformerscited by 36
- 2015Taming the Wild: A Unified Analysis of Hogwild!-Style Algorithmscited by 31
- 2020Trading Data For Learning: Incentive Mechanism For On-Device Federated Learningcited by 21
- 2021Training Multilingual Pre-trained Language Model with Byte-level Subwordscited by 16
- 2018Phonocardiographic Sensing using Deep Learning for Abnormal Heartbeat Detectioncited by 15
- 2020An Exploratory Analysis on Users' Contributions in Federated Learningcited by 15
- 2015Stop Wasting My Gradients: Practical SVRGcited by 14
- 2020Estimation of Individual Device Contributions for Incentivizing Federated Learningcited by 14
- 2014Parallel training of DNNs with Natural Gradient and Parameter Averagingcited by 11
- 2020A VCG-based Fair Incentive Mechanism for Federated Learningcited by 11
- 2020FedServing: A Federated Prediction Serving Framework Based on Incentive Mechanismcited by 10
- 2017Sequence-to-Sequence ASR Optimization via Reinforcement Learningcited by 9
- 2021Incentive Mechanism Design for Federated Learning: Hedonic Game Approachcited by 8
- 2019Generative adversarial network-based glottal waveform model for statistical parametric speech synthesiscited by 7
50 most cited of 55 All 55 (JSON)
Built on by
Citing papers per year: 2025: 1.
1 paper.