paper-with-me

홈 › Papers

Comparison of Soft and Hard Target RNN-T Distillation for Large-scale ASR

2022-10-11 · Dongseong Hwang, Khe Chai Sim, Yu Zhang, Trevor Strohman

Knowledge distillation is an effective machine learning technique to transfer knowledge from a teacher model to a smaller student model, especially with unlabeled data. In this paper, we focus on knowledge distillation for the RNN-T model, which is widely used in state-of-the-art (SoTA) automatic speech recognition (ASR). Specifically, we compared using soft and hard target distillation to train large-scaleRNN-T models on the LibriSpeech/LibriLight public dataset (60k hours) and our in-house data (600k hours). We found that hard tar-gets are more effective when the teacher and student have different architecture, such as large teacher and small streaming student. On the other hand, soft target distillation works better in self-training scenario like iterative large teacher training. For a large model with0.6B weights, we achieve a new SoTA word error rate (WER) on LibriSpeech (8% relative improvement on dev-other) using Noisy Student Training with soft target distillation. It also allows our production teacher to adapt new data domain continuously.

📄 PDF Abstract BibTeX arXiv:2210.05793

Code (0)

등록된 구현이 없습니다.

Tasks

Automatic Speech RecognitionAutomatic Speech Recognition (ASR)Knowledge Distillationspeech-recognitionSpeech RecognitionTAR

Methods 이 논문이 사용한 방법론

Stochastic Depth Stochastic Depth aims to shrink the depth of a network during training, while keeping it unchanged during testing. This is achieved by randomly dropping entire…
Knowledge Distillation A very simple way to improve the performance of almost any machine learning algorithm is to train many different models on the same data and then to average their predictions.…
RandAugment 설명 없음
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…
Noisy Student 설명 없음

Similar Papers 제목 키워드 기반

SCENE OTA-FD: Self-Centering Noncoherent Estimator for Over-the-Air Federated Distillation

2026-02-17 · Hao Chen, Zavareh Bozorgasl arxiv

We propose SCENE (Self-Centering Noncoherent Estimator), a pilot-free and phase-invariant aggregation primitive for over-the-air federated distillation (OTA-FD). Each device maps its soft-label (class-probability) vector…

Efficient Federated Learning for AIoT Applications Using Knowledge Distillation

2021-11-29 · Tian Liu, Zhiwei Ling, Jun Xia, Xin Fu 외

As a promising distributed machine learning paradigm, Federated Learning (FL) trains a central model with decentralized data without compromising user privacy, which has made it widely used by Artificial Intelligence Int…

Federated LearningKnowledge Distillation

Continuous Soft Pseudo-Labeling in ASR

2022-11-11 · Tatiana Likhomanenko, Ronan Collobert, Navdeep Jaitly, Samy Bengio

Continuous pseudo-labeling (PL) algorithms such as slimIPL have recently emerged as a powerful strategy for semi-supervised learning in speech recognition. In contrast with earlier strategies that alternated between trai…

speech-recognitionSpeech Recognition

Annealing Knowledge Distillation

2021-04-14 · EACL 2021 2 · Aref Jafari, Mehdi Rezagholizadeh, Pranav Sharma, Ali Ghodsi

Significant memory and computational requirements of large deep neural networks restrict their application on edge devices. Knowledge distillation (KD) is a prominent model compression technique for deep neural networks …

image-classificationImage ClassificationKnowledge DistillationModel Compression

Automatic adaptation of object detectors to new domains using self-training

2019-04-15 · CVPR 2019 6 · Aruni RoyChowdhury, Prithvijit Chakrabarty, Ashish Singh, SouYoung Jin 외

This work addresses the unsupervised adaptation of an existing object detector to a new target domain. We assume that a large number of unlabeled videos from this domain are readily available. We automatically obtain lab…

Domain AdaptationKnowledge DistillationPedestrian DetectionUnsupervised Domain Adaptation