Search Results for author: Zhichao Wang

Found 12 papers, 0 papers with code

High-dimensional Asymptotics of Feature Learning: How One Gradient Step Improves the Representation

no code implementations3 May 2022 Jimmy Ba, Murat A. Erdogdu, Taiji Suzuki, Zhichao Wang, Denny Wu, Greg Yang

We study the first gradient descent step on the first-layer parameters $\boldsymbol{W}$ in a two-layer neural network: $f(\boldsymbol{x}) = \frac{1}{\sqrt{N}}\boldsymbol{a}^\top\sigma(\boldsymbol{W}^\top\boldsymbol{x})$, where $\boldsymbol{W}\in\mathbb{R}^{d\times N}, \boldsymbol{a}\in\mathbb{R}^{N}$ are randomly initialized, and the training objective is the empirical MSE loss: $\frac{1}{n}\sum_{i=1}^n (f(\boldsymbol{x}_i)-y_i)^2$.

Causal Disentanglement for Semantics-Aware Intent Learning in Recommendation

no code implementations5 Feb 2022 Xiangmeng Wang, Qian Li, Dianer Yu, Peng Cui, Zhichao Wang, Guandong Xu

Traditional recommendation models trained on observational interaction data have generated large impacts in a wide range of applications, it faces bias problems that cover users' true intent and thus deteriorate the recommendation effectiveness.


IQDUBBING: Prosody modeling based on discrete self-supervised speech representation for expressive voice conversion

no code implementations2 Jan 2022 Wendong Gan, Bolong Wen, Ying Yan, Haitao Chen, Zhichao Wang, Hongqiang Du, Lei Xie, Kaixuan Guo, Hai Li

Specifically, prosody vector is first extracted from pre-trained VQ-Wav2Vec model, where rich prosody information is embedded while most speaker and environment information are removed effectively by quantization.

Quantization Voice Conversion

Multi-speaker Multi-style Text-to-speech Synthesis With Single-speaker Single-style Training Data Scenarios

no code implementations23 Dec 2021 Qicong Xie, Tao Li, Xinsheng Wang, Zhichao Wang, Lei Xie, Guoqiao Yu, Guanglu Wan

Moreover, the explicit prosody features used in the prosody predicting module can increase the diversity of synthetic speech by adjusting the value of prosody features.

Speech Synthesis Style Transfer +1

One-shot Voice Conversion For Style Transfer Based On Speaker Adaptation

no code implementations24 Nov 2021 Zhichao Wang, Qicong Xie, Tao Li, Hongqiang Du, Lei Xie, Pengcheng Zhu, Mengxiao Bi

One-shot style transfer is a challenging task, since training on one utterance makes model extremely easy to over-fit to training data and causes low speaker similarity and lack of expressiveness.

Style Transfer Voice Conversion

Deformed semicircle law and concentration of nonlinear random matrices for ultra-wide neural networks

no code implementations20 Sep 2021 Zhichao Wang, Yizhe Zhu

In this paper, we study the two-layer fully connected neural network given by $f(X)=\frac{1}{\sqrt{d_1}}\boldsymbol{a}^\top\sigma\left(WX\right)$, where $X\in\mathbb{R}^{d_0\times n}$ is a deterministic data matrix, $W\in\mathbb{R}^{d_1\times d_0}$ and $\boldsymbol{a}\in\mathbb{R}^{d_1}$ are random Gaussian weights, and $\sigma$ is a nonlinear activation function.

Enriching Source Style Transfer in Recognition-Synthesis based Non-Parallel Voice Conversion

no code implementations16 Jun 2021 Zhichao Wang, Xinyong Zhou, Fengyu Yang, Tao Li, Hongqiang Du, Lei Xie, Wendong Gan, Haitao Chen, Hai Li

Specifically, prosodic features are used to explicit model prosody, while VAE and reference encoder are used to implicitly model prosody, which take Mel spectrum and bottleneck feature as input respectively.

Style Transfer Voice Conversion

WNARS: WFST based Non-autoregressive Streaming End-to-End Speech Recognition

no code implementations8 Apr 2021 Zhichao Wang, Wenwen Yang, Pan Zhou, Wei Chen

Recently, attention-based encoder-decoder (AED) end-to-end (E2E) models have drawn more and more attention in the field of automatic speech recognition (ASR).

Automatic Speech Recognition

Spectra of the Conjugate Kernel and Neural Tangent Kernel for linear-width neural networks

no code implementations NeurIPS 2020 Zhou Fan, Zhichao Wang

We study the eigenvalue distributions of the Conjugate Kernel and Neural Tangent Kernel associated to multi-layer feedforward neural networks.

Polynomial Representation for Persistence Diagram

no code implementations CVPR 2019 Zhichao Wang, Qian Li, Gang Li, Guandong Xu

In this work, we discover a set of general polynomials that vanish on vectorized PDs and extract the task-adapted feature representation from these polynomials.

Topological Data Analysis

Cannot find the paper you are looking for? You can Submit a new open access paper.