46 citations · 153 across the 11 of their papers we have counts for
12 papers
The ISCSLP 2022 Intelligent Cockpit Speech Recognition Challenge (ICSRC): Dataset, Tracks, Baseline and Results
Ao Zhang, Fan Yu, Kaixun Huang +7
This paper summarizes the outcomes from the ISCSLP 2022 Intelligent Cockpit Speech Recognition Challenge (ICSRC). We first address the necessity of the challenge and then introduce…
The NPU-ASLP System for The ISCSLP 2022 Magichub Code-Swiching ASR Challenge
Yuhao Liang, Peikun Chen, Fan Yu +3
This paper describes our NPU-ASLP system submitted to the ISCSLP 2022 Magichub Code-Switching ASR Challenge. In this challenge, we first explore several popular end-to-end ASR arch…
A GPU-specialized Inference Parameter Server for Large-Scale Deep Recommendation Models
Yingcan Wei, Matthias Langer, Fan Yu +4
Recommendation systems are of crucial importance for a variety of modern apps and web services, such as news feeds, social networks, e-commerce, search, etc. To achieve peak predic…
Merlin HugeCTR: GPU-accelerated Recommender System Training and Inference
Joey Wang, Yingcan Wei, Minseok Lee +9
In this talk, we introduce Merlin HugeCTR. Merlin HugeCTR is an open source, GPU-accelerated integration framework for click-through rate estimation. It optimizes both training and…
BDF SAV schemes for time-fractional Allen-Cahn dissipative systems
Fan Yu, Minghua Chen
Recently, the error analysis of BDF SAV (scalar auxiliary variable) schemes are given in \cite{Huangg:20} for the classical Allen-Cahn equation. Howev…
AsymptoticNG: A regularized natural gradient optimization algorithm with look-ahead strategy
Zedong Tang, Fenlong Jiang, Junke Song +5
Optimizers that further adjust the scale of gradient, such as Adam, Natural Gradient (NG), etc., despite widely concerned and used by the community, are often found poor generaliza…
Unified Streaming and Non-streaming Two-pass End-to-end Model for Speech Recognition
Binbin Zhang, Di Wu, Zhuoyuan Yao +7
In this paper, we present a novel two-pass approach to unify streaming and non-streaming end-to-end (E2E) speech recognition in a single model. Our model adopts the hybrid CTC/atte…
Eigenvalue-corrected Natural Gradient Based on a New Approximation
Kai-Xin Gao, Xiao-Lei Liu, Zheng-Hai Huang +5
Using second-order optimization methods for training deep neural networks (DNNs) has attracted many researchers. A recently proposed method, Eigenvalue-corrected Kronecker Factoriz…
A Trace-restricted Kronecker-Factored Approximation to Natural Gradient
Kai-Xin Gao, Xiao-Lei Liu, Zheng-Hai Huang +4
Second-order optimization methods have the ability to accelerate convergence by modifying the gradient through the curvature matrix. There have been many attempts to use second-ord…
Backward difference formula: The energy technique for subdiffusion equation
Minghua Chen, Fan Yu, Zhi Zhou
Based on the equivalence of A-stability and G-stability, the energy technique of the six-step BDF method for the heat equation has been discussed in [Akrivis, Chen, Yu, Zhou, Math.…
The energy technique for the six-step BDF method
Georgios Akrivis, Minghua Chen, Fan Yu +1
In combination with the Grenander--Szegö theorem, we observe that a relaxed positivity condition on multipliers, milder than the basic %fundamental requirement of the Nevanlinna--O…
TensorOpt: Exploring the Tradeoffs in Distributed DNN Training with Auto-Parallelism
Zhenkun Cai, Kaihao Ma, Xiao Yan +5
A good parallelization strategy can significantly improve the efficiency or reduce the cost for the distributed training of deep neural networks (DNNs). Recently, several methods h…