Yuhao Liang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Yuhao Liang has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Yu-Sheng Lin
Sichuan University Chengdu 610207 China
6 shared publications - Kunye Li
School of Electronics and Information Technology, Sun Yat-sen University, Guangzhou, China
2 shared publications - Lei Xie
Speech and Language Processing Group (ASLP@NPU) Northwestern Polytechnical University Xi'an China
2 shared publications - Yu Fan
Beijing University of Posts and Telecommunications
2 shared publications - Daoye Zheng
Department of Electrical and Computer Engineering National University of Singapore Singapore Singapore
1 shared publication Hao Wen Tan
Department of Chemical Engineering and Science, Shanghai University, Shanghai, China
1 shared publication
Latest Publications
- 29 April 2025
- Published by Elsevier in Materials Chemistry and Physics
- Vol. 341
MEMS electrothermal actuator with highly linear actuation and stable characteristics
- 29 April 2025
- Published by Elsevier in Materials Chemistry and Physics
- Vol. 341
MEMS electrothermal actuator with highly linear actuation and stable characteristics
- 20 January 2025
- Published by Elsevier in Materials Today Communications
- Vol. 43
Tunable visible metasurface for color sensor application
- 20 January 2025
- Published by Elsevier in Materials Today Communications
- Vol. 43
Tunable visible metasurface for color sensor application
- Open Access
- 21 December 2024
- Published by Elsevier in Results in Engineering
- Vol. 25
Design and analysis of micro-electrostatic actuator with trapezoidal-shaped comb fingers and folded beams
- 21 December 2024
- Published by Elsevier in Results in Engineering
- Vol. 25
Open AccessDesign and analysis of micro-electrostatic actuator with trapezoidal-shaped comb fingers and folded beams
- 16 September 2024
- Published by Royal Society of Chemistry (RSC) in Journal of Materials Chemistry C
- Vol. 12 (37) , 14823-14832
Broadband polarization conversion metasurface for beam deflection with switchable characteristics
A broadband polarization conversion metasurface (PCM) for anomalous reflection and beam deflection with switchable characteristics.
- 16 September 2024
- Published by Royal Society of Chemistry (RSC) in Journal of Materials Chemistry C
- Vol. 12 (37) , 14823-14832
Broadband polarization conversion metasurface for beam deflection with switchable characteristics
A broadband polarization conversion metasurface (PCM) for anomalous reflection and beam deflection with switchable characteristics.
- Open Access
- 24 August 2024
- Published by Springer Nature in Discover Nano
- Vol. 19 (1) , 1-9
MEMS-based meta-emitter with actively tunable radiation power characteristic
We propose a meta-emitter based on micro-electro-mechanical system (MEMS) technology. The main structure of the meta-emitter unit cell is composed of four
- 24 August 2024
- Published by Springer Nature in Discover Nano
- Vol. 19 (1) , 1-9
Open AccessMEMS-based meta-emitter with actively tunable radiation power characteristic
We propose a meta-emitter based on micro-electro-mechanical system (MEMS) technology. The main structure of the meta-emitter unit cell is composed of four
- Open Access
- 16 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 88, 1-7
Sa-Paraformer: Non-Autoregressive End-To-End Speaker-Attributed ASR
Joint modeling of multi-speaker ASR and speaker diarization has recently shown promising results in speaker-attributed automatic speech recognition (SA-ASR). Although being able to obtain state-of-the-art (SOTA) performance, most of the studies are based on an autoregressive (AR) decoder which generates tokens one-by-one and results in a large real-time factor (RTF). To speed up inference, we introduce a recently proposed non-autoregressive model Paraformer as an acoustic model in the SA-ASR model. Paraformer uses a single-step decoder to enable parallel generation, obtaining comparable performance to the SOTA AR transformer models. Besides, we propose a speaker-filling strategy to reduce speaker identification errors and adopt an inter-CTC strategy to enhance the encoder’s ability in acoustic modeling. Experiments on the AliMeeting corpus show that our model outperforms the cascaded SA-ASR model by a 6.1% relative speaker-dependent character error rate (SD-CER) reduction on the test set. Moreover, our model achieves a comparable SD-CER of 34.8% with only 1/10 RTF compared with the SOTA joint AR SA-ASR model.
- 16 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 88, 1-7
Open AccessSa-Paraformer: Non-Autoregressive End-To-End Speaker-Attributed ASR
Joint modeling of multi-speaker ASR and speaker diarization has recently shown promising results in speaker-attributed automatic speech recognition (SA-ASR). Although being able to obtain state-of-the-art (SOTA) performance, most of the studies are based on an autoregressive (AR) decoder which generates tokens one-by-one and results in a large real-time factor (RTF). To speed up inference, we introduce a recently proposed non-autoregressive model Paraformer as an acoustic model in the SA-ASR model. Paraformer uses a single-step decoder to enable parallel generation, obtaining comparable performance to the SOTA AR transformer models. Besides, we propose a speaker-filling strategy to reduce speaker identification errors and adopt an inter-CTC strategy to enhance the encoder’s ability in acoustic modeling. Experiments on the AliMeeting corpus show that our model outperforms the cascaded SA-ASR model by a 6.1% relative speaker-dependent character error rate (SD-CER) reduction on the test set. Moreover, our model achieves a comparable SD-CER of 34.8% with only 1/10 RTF compared with the SOTA joint AR SA-ASR model.
- Open Access
- 16 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. abs/1211.3, 1-7
BA-MoE: Boundary-Aware Mixture-of-Experts Adapter for Code-Switching Speech Recognition
Mixture-of-experts based models, which use language experts to extract language-specific representations effectively, have been well applied in code-switching automatic speech recognition. However, there is still substantial space to improve as similar pronunciation across languages may result in ineffective multi-language modeling and inaccurate language boundary estimation. To eliminate these drawbacks, we propose a cross-layer language adapter and a boundary-aware training method, namely Boundary-Aware Mixture-of-Experts (BA-MoE). Specifically, we introduce language-specific adapters to separate language-specific representations and a unified gating layer to fuse representations within each encoder layer. Second, we compute language adaptation loss of the mean output of each language-specific adapter to improve the adapter module’s language-specific representation learning. Besides, we utilize a boundary-aware predictor to learn boundary representations for dealing with language boundary confusion. Our approach achieves significant performance improvement, reducing the mixture error rate by 16.55% compared to the baseline on the ASRU 2019 Mandarin-English code-switching challenge dataset.
- 16 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. abs/1211.3, 1-7
Open AccessBA-MoE: Boundary-Aware Mixture-of-Experts Adapter for Code-Switching Speech Recognition
Mixture-of-experts based models, which use language experts to extract language-specific representations effectively, have been well applied in code-switching automatic speech recognition. However, there is still substantial space to improve as similar pronunciation across languages may result in ineffective multi-language modeling and inaccurate language boundary estimation. To eliminate these drawbacks, we propose a cross-layer language adapter and a boundary-aware training method, namely Boundary-Aware Mixture-of-Experts (BA-MoE). Specifically, we introduce language-specific adapters to separate language-specific representations and a unified gating layer to fuse representations within each encoder layer. Second, we compute language adaptation loss of the mean output of each language-specific adapter to improve the adapter module’s language-specific representation learning. Besides, we utilize a boundary-aware predictor to learn boundary representations for dealing with language boundary confusion. Our approach achieves significant performance improvement, reducing the mixture error rate by 16.55% compared to the baseline on the ASRU 2019 Mandarin-English code-switching challenge dataset.
- 11 October 2022
- Published by Taylor & Francis
- p. 671-679
Vehicle path planning of intelligent vehicle-road system based on vehicle historical path trajectory
With the development of society and the improvement of national economy, the number of vehicles in the road system is increasing, and the problem of road traffic congestion is becoming more severe. Hence, reasonable planning of vehicle routing is becoming more important. The vehicle path planning of intelligent vehicle road system not only needs to solve the current traffic congestion problem, but also needs to meet the dynamic travel needs of residents. By processing the important traffic element information of vehicle historical path trajectory, this paper extracts the effective content of the trajectory, and assigns dynamic weight to each section. Based on A* algorithm and Yen's algorithm, an improved A* algorithm is innovatively proposed. According to the traffic flow in different periods, k optimal roads are recommended. For given constraints, these roads will be filtered, and the filtering results will be more consistent. The actual traffic conditions can effectively alleviate the traffic pressure and provide convenience for traffic travelers to choose appropriate travel roads.
- 11 October 2022
- Published by Taylor & Francis
- p. 671-679
Vehicle path planning of intelligent vehicle-road system based on vehicle historical path trajectory
With the development of society and the improvement of national economy, the number of vehicles in the road system is increasing, and the problem of road traffic congestion is becoming more severe. Hence, reasonable planning of vehicle routing is becoming more important. The vehicle path planning of intelligent vehicle road system not only needs to solve the current traffic congestion problem, but also needs to meet the dynamic travel needs of residents. By processing the important traffic element information of vehicle historical path trajectory, this paper extracts the effective content of the trajectory, and assigns dynamic weight to each section. Based on A* algorithm and Yen's algorithm, an improved A* algorithm is innovatively proposed. According to the traffic flow in different periods, k optimal roads are recommended. For given constraints, these roads will be filtered, and the filtering results will be more consistent. The actual traffic conditions can effectively alleviate the traffic pressure and provide convenience for traffic travelers to choose appropriate travel roads.
- 1 November 2021
- Published by Pleiades Publishing Ltd in Polymer Science, Series A
- Vol. 63 (6) , 828-841
Synthesis of Nano-polypyrrole/Reduced Graphene Oxide via Double Emulsion Method
The nano-polypyrrole/reduced graphene oxide composites are successfully prepared through double emulsion method (water in oil in water system). Graphene oxide is synthesized by the modified Hummers method, and reduced graphene oxide is obtained via thermal reduction method. Effects of the conditions during polymerization on the morphology of the polypyrrole nanospheres have been investigated. Fourier transform infrared spectroscopy (FTIR), X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM) and electrochemical workstation (CHI 660E) are utilized to characterize the composites. The synthesized composites have the specific capacitance of 855 F/g at a scan rate of 1 mV/s and the specific capacitance retention can still maintain 88% after 800 cycles, which indicates good electrochemical performance and good cyclability.
- 1 November 2021
- Published by Pleiades Publishing Ltd in Polymer Science, Series A
- Vol. 63 (6) , 828-841
Synthesis of Nano-polypyrrole/Reduced Graphene Oxide via Double Emulsion Method
The nano-polypyrrole/reduced graphene oxide composites are successfully prepared through double emulsion method (water in oil in water system). Graphene oxide is synthesized by the modified Hummers method, and reduced graphene oxide is obtained via thermal reduction method. Effects of the conditions during polymerization on the morphology of the polypyrrole nanospheres have been investigated. Fourier transform infrared spectroscopy (FTIR), X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM) and electrochemical workstation (CHI 660E) are utilized to characterize the composites. The synthesized composites have the specific capacitance of 855 F/g at a scan rate of 1 mV/s and the specific capacitance retention can still maintain 88% after 800 cycles, which indicates good electrochemical performance and good cyclability.