Xiaoyu Song
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
Research Subjects
Research Interests
Xiaoyu Song has a total of 116 co-authors. Here are the top co-authors by number of shared publications:
- Zhuowei Wang
Guangdong University of Technology, Guangzhou, China
9 shared publications - Jian-Tao Zhou
College of Computer Science, Inner Mongolia University, Hohhot, China
6 shared publications - Zonghui Li
Beijing Jiaotong University
6 shared publications - Hai Wan
Tsinghua University, Beijing, China
5 shared publications - Min Zhou
Tsinghua University
4 shared publications - Xibin Zhao
Tsinghua University, Beijing, China
4 shared publications
Latest Publications
- 29 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computers
- Vol. 75 (8) , 2769-2782
DTRRS: A Dynamic Task Real-Time Reconfiguration and Scheduling Architecture for FPGAs in Embedded Systems
Field-programmable gate arrays (FPGAs) can provide high-performance computing power for embedded systems. However, developing FPGA programs is challenging and timeconsuming. This results in FPGAs not being rapidly programmable in the field to meet variable requirements. Many programming language and architecture abstraction methods have been proposed. Although these methods address development challenges, they still do not resolve the issue of time-consuming compilation, which results in low efficiency during early-stage solution verification. To this end, this paper proposes DTRRS, an architecture designed to facilitate the rapid reconfiguration and scheduling of tasks on FPGAs. DTRRS provides programming language abstraction that supports task development using the ladder diagram (LD) programming language commonly employed in industrial programmable logic controllers (PLCs). It also provides architecture abstraction that supports the predefinition of a basic function module (BFM) set and enables task changes by dynamically establishing connections among BFMs in this set. Experimental results indicate that when the on-board function library covers the new task, DTRRS achieves a task reconfiguration time of about 0.001 seconds, significantly lower than that of the other solutions.
- 29 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computers
- Vol. 75 (8) , 2769-2782
DTRRS: A Dynamic Task Real-Time Reconfiguration and Scheduling Architecture for FPGAs in Embedded Systems
Field-programmable gate arrays (FPGAs) can provide high-performance computing power for embedded systems. However, developing FPGA programs is challenging and timeconsuming. This results in FPGAs not being rapidly programmable in the field to meet variable requirements. Many programming language and architecture abstraction methods have been proposed. Although these methods address development challenges, they still do not resolve the issue of time-consuming compilation, which results in low efficiency during early-stage solution verification. To this end, this paper proposes DTRRS, an architecture designed to facilitate the rapid reconfiguration and scheduling of tasks on FPGAs. DTRRS provides programming language abstraction that supports task development using the ladder diagram (LD) programming language commonly employed in industrial programmable logic controllers (PLCs). It also provides architecture abstraction that supports the predefinition of a basic function module (BFM) set and enables task changes by dynamically establishing connections among BFMs in this set. Experimental results indicate that when the on-board function library covers the new task, DTRRS achieves a task reconfiguration time of about 0.001 seconds, significantly lower than that of the other solutions.
- 20 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (10) , 20822-20835
Bidirectional-Graph Attention Networks Parallel Encoder for Data Imputation and Fault Diagnosis of Industrial Robots
Safe operation is a key concern for industrial robots. However, due to hardware failures and unstable data transmission issues, the multivariate time-series data generated by these axes often contain missing or corrupted signals, which severely hinders downstream tasks such as fault diagnosis. In addition, the substantial volume of industrial data demands considerable time for training time-series imputation models and subsequent classification models. To address these challenges, this study proposes a multitask approach that serves both the data imputation and fault diagnosis tasks for industrial robots. In particular, the parameters trained in the imputation model can be transferred to the fault diagnosis model, enhancing its performance and efficiency. A multitask method named bidirectional-graph attention networks parallel encoder (Bi-GATPE) is proposed, which employs a bidirectional-graph attention network to capture the spatial dependencies among the various variables of industrial robots. Subsequently, a parallel encoder with diagonal-filter attention is designed to model temporal correlations. This dual approach improves the accuracy and training speed for both the imputation and fault diagnosis tasks. The experimental studies based on real industrial robot datasets demonstrate that by modifying the feature fusion layer of the imputation task and sharing the trained parameters with the fault diagnosis task, the proposed method significantly accelerates the convergence of the fault diagnosis model while also improving diagnostic accuracy. The experiments also indicate that our method shows merits in the imputation and fault diagnosis tasks. The source code of Bi-GATPE is available at: https://github.com/miten073/Bi-GATPE
- 20 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (10) , 20822-20835
Bidirectional-Graph Attention Networks Parallel Encoder for Data Imputation and Fault Diagnosis of Industrial Robots
Safe operation is a key concern for industrial robots. However, due to hardware failures and unstable data transmission issues, the multivariate time-series data generated by these axes often contain missing or corrupted signals, which severely hinders downstream tasks such as fault diagnosis. In addition, the substantial volume of industrial data demands considerable time for training time-series imputation models and subsequent classification models. To address these challenges, this study proposes a multitask approach that serves both the data imputation and fault diagnosis tasks for industrial robots. In particular, the parameters trained in the imputation model can be transferred to the fault diagnosis model, enhancing its performance and efficiency. A multitask method named bidirectional-graph attention networks parallel encoder (Bi-GATPE) is proposed, which employs a bidirectional-graph attention network to capture the spatial dependencies among the various variables of industrial robots. Subsequently, a parallel encoder with diagonal-filter attention is designed to model temporal correlations. This dual approach improves the accuracy and training speed for both the imputation and fault diagnosis tasks. The experimental studies based on real industrial robot datasets demonstrate that by modifying the feature fusion layer of the imputation task and sharing the trained parameters with the fault diagnosis task, the proposed method significantly accelerates the convergence of the fault diagnosis model while also improving diagnostic accuracy. The experiments also indicate that our method shows merits in the imputation and fault diagnosis tasks. The source code of Bi-GATPE is available at: https://github.com/miten073/Bi-GATPE
- 3 February 2026
- Published by IOP Publishing in Engineering Research Express
- Vol. 8 (4)
WIFFormer: Wavelet-Based Illumination Fusion Transformer for Low-light Image Enhancement
Low-light image enhancement aims to restore the lighting and content features of captured images in dark environments, which is crucial to improving visual experience and supporting advanced computer vision tasks. The existing low-light image enhancement methods based on the Convolutional Neural Network have poor performance in global color restoration and can not effectively deal with the problem of uneven illumination. In recent years, although the low-light image enhancement method based on the Vision Transformer (ViT) improves the global color quality of image restoration, it often ignores the processing of details, and the computational efficiency needs to be improved. To address the above problems, we propose WIFFormer, a low-light image enhancement method based on Transformer structures and combined with two-dimensional Discrete Wavelet Transform (2D-DWT) to balance global color restoration and detail preservation. Specifically, WIFFormer consists of two core components: Wavelet Illuminator (WI) and Wavelet Content-Detail Restorer (WR). WI generates a preliminary illumination image, while WR restores global color content and details on the illuminated image. In WR, we design a double-branch structure based on 2D-DWT, where the Illumination Fusion Branch (IFB) adopts a UNet network equipped with the Illumination Fusion Recovery Block (IFRB) to recover low-frequency content information and the Detail Reconstruction Branch (DRB) reconstructs high-frequency detail information from three directions. Extensive experiments on five low-light image enhancement benchmarks demonstrate the advantages of WIFFormer in image restoration quality and efficiency compared to existing State-Of-The-Art (SOTA) methods. In addition, the experimental results of the downstream advanced computer vision task also reveal the potential value of our method in the fields of automatic driving and security monitoring in the dark.
- 3 February 2026
- Published by IOP Publishing in Engineering Research Express
- Vol. 8 (4)
WIFFormer: Wavelet-Based Illumination Fusion Transformer for Low-light Image Enhancement
Low-light image enhancement aims to restore the lighting and content features of captured images in dark environments, which is crucial to improving visual experience and supporting advanced computer vision tasks. The existing low-light image enhancement methods based on the Convolutional Neural Network have poor performance in global color restoration and can not effectively deal with the problem of uneven illumination. In recent years, although the low-light image enhancement method based on the Vision Transformer (ViT) improves the global color quality of image restoration, it often ignores the processing of details, and the computational efficiency needs to be improved. To address the above problems, we propose WIFFormer, a low-light image enhancement method based on Transformer structures and combined with two-dimensional Discrete Wavelet Transform (2D-DWT) to balance global color restoration and detail preservation. Specifically, WIFFormer consists of two core components: Wavelet Illuminator (WI) and Wavelet Content-Detail Restorer (WR). WI generates a preliminary illumination image, while WR restores global color content and details on the illuminated image. In WR, we design a double-branch structure based on 2D-DWT, where the Illumination Fusion Branch (IFB) adopts a UNet network equipped with the Illumination Fusion Recovery Block (IFRB) to recover low-frequency content information and the Detail Reconstruction Branch (DRB) reconstructs high-frequency detail information from three directions. Extensive experiments on five low-light image enhancement benchmarks demonstrate the advantages of WIFFormer in image restoration quality and efficiency compared to existing State-Of-The-Art (SOTA) methods. In addition, the experimental results of the downstream advanced computer vision task also reveal the potential value of our method in the fields of automatic driving and security monitoring in the dark.
- Open Access
Solving Segment Display Problems Using Quantum Grover's Search Algorithm
This paper introduces a new Boolean-based methodology for constructing Segment Display Problems (SDPs) in the quantum domain and solving them using Grover's quantum search algorithm. In the classical domain, the SDPs are typically solved using various techniques, such as human deduction, heuristic search, and methods for solving Boolean satisfiability (SAT) and constraint satisfaction problems (CSPs) that are based on different problem design models. In this paper, our newly introduced methodology proposes a quantum-based approach for solving such SDPs, by building their quantum oracle using binary reversible circuits and our previously proposed step-decreasing structures shaped operators (Stesso). To demonstrate the usability of this proposed method, we experimentally solve an SDP instance of the matchstick problem using Grover's algorithm with a noisy simulated quantum computer implemented in Qiskit.
Open AccessSolving Segment Display Problems Using Quantum Grover's Search Algorithm
This paper introduces a new Boolean-based methodology for constructing Segment Display Problems (SDPs) in the quantum domain and solving them using Grover's quantum search algorithm. In the classical domain, the SDPs are typically solved using various techniques, such as human deduction, heuristic search, and methods for solving Boolean satisfiability (SAT) and constraint satisfaction problems (CSPs) that are based on different problem design models. In this paper, our newly introduced methodology proposes a quantum-based approach for solving such SDPs, by building their quantum oracle using binary reversible circuits and our previously proposed step-decreasing structures shaped operators (Stesso). To demonstrate the usability of this proposed method, we experimentally solve an SDP instance of the matchstick problem using Grover's algorithm with a noisy simulated quantum computer implemented in Qiskit.
- 28 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Selected Areas in Communications
- Vol. PP (07338716) , 1
Time Synchronization for 5G and TSN Integrated Networking
Emerging industrial applications involving robotic collaborative operations and mobile robots require a more reliable and precise wireless network for deterministic data transmission. To meet this demand, the 3rd Generation Partnership Project (3GPP) is promoting the integration of 5th Generation Mobile Communication Technology (5G) and Time-Sensitive Networking (TSN). Time synchronization is essential for deterministic data transmission. Based on the 3GPP’s vision of the 5G and TSN integrated networking with interoperability, we improve the time synchronization of TSN to conquer the multi-gNB competition, re-transmission, and mobility problems for the integrated 5G time synchronization. We implemented the improvement mechanisms and systematically validated the performance of 5G+TSN time synchronization. Based on the simulation in 500m x 500m industrial environments, the improved time synchronization achieved a precision of 1 microsecond with interoperability between 5G nodes and TSN nodes.
- 28 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Selected Areas in Communications
- Vol. PP (07338716) , 1
Time Synchronization for 5G and TSN Integrated Networking
Emerging industrial applications involving robotic collaborative operations and mobile robots require a more reliable and precise wireless network for deterministic data transmission. To meet this demand, the 3rd Generation Partnership Project (3GPP) is promoting the integration of 5th Generation Mobile Communication Technology (5G) and Time-Sensitive Networking (TSN). Time synchronization is essential for deterministic data transmission. Based on the 3GPP’s vision of the 5G and TSN integrated networking with interoperability, we improve the time synchronization of TSN to conquer the multi-gNB competition, re-transmission, and mobility problems for the integrated 5G time synchronization. We implemented the improvement mechanisms and systematically validated the performance of 5G+TSN time synchronization. Based on the simulation in 500m x 500m industrial environments, the improved time synchronization achieved a precision of 1 microsecond with interoperability between 5G nodes and TSN nodes.
- 15 April 2025
- Published by Springer Nature in Data Mining and Knowledge Discovery
- Vol. 39 (3) , 1-27
HyEED: embedding learning of knowledge graphs with entity description in hyperbolic space
- 15 April 2025
- Published by Springer Nature in Data Mining and Knowledge Discovery
- Vol. 39 (3) , 1-27
HyEED: embedding learning of knowledge graphs with entity description in hyperbolic space
- Open Access
- 3 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 62739-62752
An Effective Genetic Algorithm for Mixed Precision
The precision of floating-point numbers is a critical task in high-performance computing. Many scientific applications rely on floating-point arithmetic, but excessive precision can lead to unnecessary computational overhead. Reducing precision may introduce unacceptable errors. Addressing this trade-off is essential for optimizing performance while ensuring numerical accuracy. In this paper, we present a genetic algorithm-based approach for tuning the precision of floating-point computations. Our method leverages algorithmic differentiation and first-order Taylor series approximation to assess the impact of precision variations efficiently. We employ stochastic partitioning algorithms with multiple precision combinations that meet the error requirements. Moreover, we present a genetic heuristic algorithm to determine the maximum number of variables that can sustain precision alterations without compromising the desired error threshold. The proposed approach is evaluated across various benchmark programs, analyzing the effects of precision tuning under increasing error thresholds. Our findings reveal that, for a majority of these programs, reducing precision through partitioning leads to significant performance enhancements, with improvements of up to 15%.
- 3 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 62739-62752
Open AccessAn Effective Genetic Algorithm for Mixed Precision
The precision of floating-point numbers is a critical task in high-performance computing. Many scientific applications rely on floating-point arithmetic, but excessive precision can lead to unnecessary computational overhead. Reducing precision may introduce unacceptable errors. Addressing this trade-off is essential for optimizing performance while ensuring numerical accuracy. In this paper, we present a genetic algorithm-based approach for tuning the precision of floating-point computations. Our method leverages algorithmic differentiation and first-order Taylor series approximation to assess the impact of precision variations efficiently. We employ stochastic partitioning algorithms with multiple precision combinations that meet the error requirements. Moreover, we present a genetic heuristic algorithm to determine the maximum number of variables that can sustain precision alterations without compromising the desired error threshold. The proposed approach is evaluated across various benchmark programs, analyzing the effects of precision tuning under increasing error thresholds. Our findings reveal that, for a majority of these programs, reducing precision through partitioning leads to significant performance enhancements, with improvements of up to 15%.
- Open Access
- 20 February 2025
- Published by Walter de Gruyter GmbH in Quantum Information & Computation
- Vol. 25 (1) , 57-70
Quantum Finite Automaton Using Ternary Rotation Quantum Gates and Chrestenson Family Quantum Gates
Quantum automata can solve certain problems with a smaller state space than classical automata. We developed a quantum finite automaton using ternary rotation quantum gates and the Chrestenson family of ternary quantum gates. The main idea of this paper is to show how to combine rotation ternary quantum circuit-based Quantum Finite Automaton and quantum reversible circuit-based Deterministic Finite Automaton to build a more powerful machine. The combined machine can enable more complex language and pattern recognition. The developed quantum finite automaton and resulting combined machine can be used for robotics applications such as language, gesture, and motion recognition.
- 20 February 2025
- Published by Walter de Gruyter GmbH in Quantum Information & Computation
- Vol. 25 (1) , 57-70
Open AccessQuantum Finite Automaton Using Ternary Rotation Quantum Gates and Chrestenson Family Quantum Gates
Quantum automata can solve certain problems with a smaller state space than classical automata. We developed a quantum finite automaton using ternary rotation quantum gates and the Chrestenson family of ternary quantum gates. The main idea of this paper is to show how to combine rotation ternary quantum circuit-based Quantum Finite Automaton and quantum reversible circuit-based Deterministic Finite Automaton to build a more powerful machine. The combined machine can enable more complex language and pattern recognition. The developed quantum finite automaton and resulting combined machine can be used for robotics applications such as language, gesture, and motion recognition.
- 11 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 44 (6) , 2328-2342
A Reduced State-Space Generation Method for Concurrent Systems Based on CPN-PR Model
Colored Petri Nets (CPNs) provide descriptions of the concurrent behaviors for software and hardware. Model checking based on CPNs is an effective method to simulate and verify the concurrent behavior in system design. However, the model checking method traverses the full state space, which suffers from the state space explosion problem. A reduced state space generation method related to the property of concurrent systems is proposed. Specifically, we extend CPNs to define a property-related model (CPN-PR) and give a property-related analysis method whose results can be used to generate the CPN-PR model. A reduced state space generation method is developed based on enabled binding element filtering rules. The stutter trace equivalence between the state spaces of CPN and CPN-PR has been proven by showing that the reduced state space may not change the model checking result. A comparison experiment is conducted to demonstrate the effectiveness of our method.
- 11 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 44 (6) , 2328-2342
A Reduced State-Space Generation Method for Concurrent Systems Based on CPN-PR Model
Colored Petri Nets (CPNs) provide descriptions of the concurrent behaviors for software and hardware. Model checking based on CPNs is an effective method to simulate and verify the concurrent behavior in system design. However, the model checking method traverses the full state space, which suffers from the state space explosion problem. A reduced state space generation method related to the property of concurrent systems is proposed. Specifically, we extend CPNs to define a property-related model (CPN-PR) and give a property-related analysis method whose results can be used to generate the CPN-PR model. A reduced state space generation method is developed based on enabled binding element filtering rules. The stutter trace equivalence between the state spaces of CPN and CPN-PR has been proven by showing that the reduced state space may not change the model checking result. A comparison experiment is conducted to demonstrate the effectiveness of our method.
- 3 October 2024
- Published by Springer Nature in Knowledge and Information Systems
- Vol. 67 (1) , 661-687
HGCGE: hyperbolic graph convolutional networks-based knowledge graph embedding for link prediction
- 3 October 2024
- Published by Springer Nature in Knowledge and Information Systems
- Vol. 67 (1) , 661-687
HGCGE: hyperbolic graph convolutional networks-based knowledge graph embedding for link prediction