Jinyu Li

ORCID: 0000-0003-3045-6220
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Ion-surface interactions and analysis
  • Integrated Circuits and Semiconductor Failure Analysis
  • Particle accelerators and beam dynamics
  • Diamond and Carbon-based Materials Research
  • Speech and Audio Processing
  • Solar Thermal and Photovoltaic Systems
  • Plasma Diagnostics and Applications
  • Particle Accelerators and Free-Electron Lasers
  • Speech Recognition and Synthesis
  • Electron and X-Ray Spectroscopy Techniques
  • Blind Source Separation Techniques
  • Solar-Powered Water Purification Methods
  • Yersinia bacterium, plague, ectoparasites research
  • Advanced Adaptive Filtering Techniques
  • Speech and dialogue systems
  • solar cell performance optimization
  • Vector-borne infectious diseases
  • Face and Expression Recognition
  • Bee Products Chemical Analysis
  • Optical Coatings and Gratings
  • Magnetic confinement fusion research
  • Electrostatic Discharge in Electronics
  • Near-Field Optical Microscopy
  • Lubricants and Their Additives
  • Chemical Looping and Thermochemical Processes

Jilin Agricultural Science and Technology University
2025

Fuzhou University
2025

China University of Petroleum, East China
2023-2024

National Institute for Communicable Disease Control and Prevention
2024

Shandong University
2024

National Synchrotron Radiation Laboratory
2018-2023

University of Science and Technology of China
2004-2023

Chinese Academy of Sciences
1998-2019

Institute of Modern Physics
1998-2019

Beijing Institute of Technology
2017

10.1109/icassp49660.2025.10889780 article EN ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2025-03-12

Recently, zero-shot text-to-speech (TTS) systems, capable of synthesizing any speaker's voice from a short audio prompt, have made rapid advancements. However, the quality generated speech significantly deteriorates when prompt contains noise, and limited research has been conducted to address this issue. In paper, we explored various strategies enhance noisy prompts within context flow-matching-based TTS. Our investigation includes comprehensive training strategies: unsupervised...

10.21437/interspeech.2024-1336 article EN Interspeech 2022 2024-09-01

The energy dependence of nanostructures on a photoresist produced by ion bombardment (IB) under normal incidence is studied through atomic force microscopy and time-of-flight secondary mass spectroscopy (ToF-SIMS). energy-dependent morphology evolved from weak islands via nanoholes to smooth surface the resist; in particular, were for broad range 300–550 eV. enrichment light components layer irradiated resist was illustrated owing strong decomposition IB. morphologies explained according...

10.1016/j.apsusc.2020.146510 article EN cc-by-nc-nd Applied Surface Science 2020-04-28

10.1016/j.nimb.2011.12.041 article EN Nuclear Instruments and Methods in Physics Research Section B Beam Interactions with Materials and Atoms 2011-12-30

Random nanohole morphologies on a photoresist surface were produced spontaneously by ion bombardment, from normal to near-normal incidence. At an optimized energy, the critical parameters (diameter and depth) uniformity of these nanoholes can be tailored with multi-ion parameters, including fluence (i.e. bombardment time), flux, incidence angle. The temporal evolution structures, at incidence, showed growth stabilization regimes. mass redistribution effect distribution energy deposition led...

10.1016/j.apsusc.2020.147515 article EN cc-by-nc-nd Applied Surface Science 2020-08-13

In this paper, the minimum classification error (MCE) method is extended to optimize both linear discriminant analysis (LDA) transformation and parameters for dimensionality reduction. Firstly, under HMM-based continuous speech recognition (CSR) framework, we use MCE criterion conventional reduction method, which uses LDA transform standard MFCC. Then, a new proposed. combination of discrete cosine (DCT) LDA, as used in replaced by single transformation, optimized according along with...

10.1109/icassp.2004.1325941 article EN IEEE International Conference on Acoustics Speech and Signal Processing 2004-09-28

The morphology evolution of self-organized nanopatterns induced during Ar+ ion bombardment (IB) with Mo co-deposition on fused silica (SiO2) surfaces at different incidence angles and fluences was investigated by using atomic force microscopy transmission electron microscopy. For pure IB from 30° to 70°, SiO2 evolve being flat, via ripples, direction-transversed ripples. In contrast, the same fluence angles, simultaneous leads significant terraced structures significantly enhanced roughness...

10.1063/1.5039565 article EN cc-by Applied Physics Letters 2018-07-16

Abstract Ion bombardment (IB) is a promising nanofabrication technique for producing nanoripples. A critical issue that restricts the application of IB limited quality IB-induced Photoresist (PR) and antireflection coating (ARC) are technological relevance lithographic exposure processes. Moreover, to improve self-organized nanoripples, in this study, PR/ARC bilayer was bombarded at an incidence angle 50°. The surface normalized defect density power spectral density, obtained via scanning...

10.1088/1361-6528/ac074e article EN cc-by Nanotechnology 2021-06-02

Ion bombardment (IB) is a promising nanofabrication tool for self-organized nanostructures. When ions bombard nominally flat solid surface, nanoripples can be induced on the irradiated target which are called intrinsic of material. The degree ordering an outstanding issue to overcome, similar other self-organization methods. In this study, IB-induced bilayer systems with enhanced quality revisited from perspective guided self-organization. First, power spectral density (PSD) entropy...

10.3390/e25010170 article EN cc-by Entropy 2023-01-14

Recently, zero-shot text-to-speech (TTS) systems, capable of synthesizing any speaker's voice from a short audio prompt, have made rapid advancements. However, the quality generated speech significantly deteriorates when prompt contains noise, and limited research has been conducted to address this issue. In paper, we explored various strategies enhance noisy prompts within context flow-matching-based TTS. Our investigation includes comprehensive training strategies: unsupervised...

10.48550/arxiv.2406.05699 preprint EN arXiv (Cornell University) 2024-06-09

Climatic factors are closely associated with the occurrence of vector-borne diseases, and they also influence distribution vectors. The plague is related to population dynamics fleas their host animals, as well climatic conditions. This study focused on Xilingol League, utilizing flea index data from 2012 2021. Spearman correlation "Boruta" importance analysis were conducted screen for variables. A generalized additive model (GAM) was employed investigate rodent density index. GAM revealed...

10.1016/j.bsheal.2024.07.004 article EN cc-by-nc-nd Biosafety and Health 2024-07-10

Generative Error Correction (GEC) has emerged as a powerful post-processing method to enhance the performance of Automatic Speech Recognition (ASR) systems. However, we show that GEC models struggle generalize beyond specific types errors encountered during training, limiting their ability correct new, unseen at test time, particularly in out-of-domain (OOD) scenarios. This phenomenon amplifies with named entities (NEs), where, addition insufficient contextual information or knowledge about...

10.48550/arxiv.2410.13198 preprint EN arXiv (Cornell University) 2024-10-17

In this paper, we introduce V2SFlow, a novel Video-to-Speech (V2S) framework designed to generate natural and intelligible speech directly from silent talking face videos. While recent V2S systems have shown promising results on constrained datasets with limited speakers vocabularies, their performance often degrades real-world, unconstrained due the inherent variability complexity of signals. To address these challenges, decompose signal into manageable subspaces (content, pitch, speaker...

10.48550/arxiv.2411.19486 preprint EN arXiv (Cornell University) 2024-11-29

With the increasing in laser power, it is necessary to prepare protective coatings for components and devices under high power irradiation environment. Yttria-stabilized zirconia has melting point, good stability low thermal conductivity, often used as barrier coating material. However, study on its anti CW performance rare. In this paper, yttria-stabilized was deposited using atmospheric plasma spraying. The spectrophotometer, X-ray diffraction, scanning electron microscope energy...

10.1117/12.2271764 article EN Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE 2017-08-01

Abstract Pulex simulans and Polygenis gwyni are vectors of many flea‐borne diseases. They were widely recorded in the United States Mexico between 1970 2000. Maximum entropy models used to explore habitats both fleas under different climate scenarios provide scientific basis for surveillance control We screened variables by principal component analysis Pearson's correlation test evaluated model performance ROC curve. ArcMap was visualize expressions. Under current climatic conditions, medium...

10.1002/ece3.11621 article EN cc-by Ecology and Evolution 2024-07-01

A beam-profile monitor has been constructed based on a two-dimensional cross-connected-pixels anode and 128-channel picoammeter system. It can provide the total beam current, as well current projections in x- y-directions with spatial resolution of ∼1 mm. is suitable for diagnosis low-energy charged-particle beams ranging from subpicoamperes to nanoamperes, e.g., ion extracted an electron source.

10.1063/1.5094090 article EN Review of Scientific Instruments 2019-07-01
Coming Soon ...