Hao Zhang
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
Hao Zhang has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
- Haibo Zhang
School of Computing, University of Otago, Dunedin, New Zealand
4 shared publications - Zhiyi Huang
School of Computing, University of Otago, Dunedin, New Zealand
4 shared publications - Rong Hua
College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao, Shandong, China
3 shared publications - Xiulong Yi
College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao, Shandong, China
3 shared publications - You Fu
College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao, Shandong, China
3 shared publications - Chengpeng Xia
Department of Computer Science, University of Otago, Dunedin, New Zealand
2 shared publications
Latest Publications
- 12 September 2025
- Published by Elsevier in Neural Networks
- Vol. 194, 108102
Radiology report generation via visual-semantic ambivalence-aware network and focal self-critical sequence training
- 12 September 2025
- Published by Elsevier in Neural Networks
- Vol. 194, 108102
Radiology report generation via visual-semantic ambivalence-aware network and focal self-critical sequence training
- Open Access
- 8 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 1020-1033
ROCKET: An RNS-based Photonic Accelerator for High-Precision and Energy-Efficient DNN Training
- 8 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 1020-1033
Open AccessROCKET: An RNS-based Photonic Accelerator for High-Precision and Energy-Efficient DNN Training
- 27 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 44 (3) , 1494-1504
LHR-RFL: Linear Hybrid-Reward-Based Reinforced Focal Learning for Automatic Radiology Report Generation
Radiology report generation that aims to accurately describe medical findings for given images, is pivotal in contemporary computer-aided diagnosis. Recently, despite considerable progress, current radiology report generation models still struggled to achieve consistent quality across difficult and easy samples, which dramatically impacts their clinical value. To solve this problem, we explore the difficult samples mining in radiology report generation and propose the Linear Hybrid-Reward based Reinforced Focal Learning (LHR-RFL) to effectively guide the model to allocate more attention towards some difficult samples, thereby enhancing its overall performance in both general and intricate scenarios. In implementation, we first propose the Linear Hybrid-Reward (LHR) module to better quantify the learning difficulty, which employs a linear weighting scheme that assigns varying weights to three representative Natural Language Generation (NLG) evaluation metrics. Then, we propose the Reinforced Focal Learning (RFL) to adaptively adjust the contributions of difficult samples during training, thereby augmenting their impact on model optimization. The experimental results demonstrate that our proposed LHR-RFL improves the performance of the base model across all NLG evaluation metrics, achieving an average performance improvement of 20.9% and 13.2% on IU X-ray and MIMIC-CXR datasets, respectively. Further analysis also proves that our LHR-RFL can dramatically improve the quality of reports for difficult samples. The source code will be available at https://github.com/SKD-HPC/LHR-RFL.
- 27 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 44 (3) , 1494-1504
LHR-RFL: Linear Hybrid-Reward-Based Reinforced Focal Learning for Automatic Radiology Report Generation
Radiology report generation that aims to accurately describe medical findings for given images, is pivotal in contemporary computer-aided diagnosis. Recently, despite considerable progress, current radiology report generation models still struggled to achieve consistent quality across difficult and easy samples, which dramatically impacts their clinical value. To solve this problem, we explore the difficult samples mining in radiology report generation and propose the Linear Hybrid-Reward based Reinforced Focal Learning (LHR-RFL) to effectively guide the model to allocate more attention towards some difficult samples, thereby enhancing its overall performance in both general and intricate scenarios. In implementation, we first propose the Linear Hybrid-Reward (LHR) module to better quantify the learning difficulty, which employs a linear weighting scheme that assigns varying weights to three representative Natural Language Generation (NLG) evaluation metrics. Then, we propose the Reinforced Focal Learning (RFL) to adaptively adjust the contributions of difficult samples during training, thereby augmenting their impact on model optimization. The experimental results demonstrate that our proposed LHR-RFL improves the performance of the base model across all NLG evaluation metrics, achieving an average performance improvement of 20.9% and 13.2% on IU X-ray and MIMIC-CXR datasets, respectively. Further analysis also proves that our LHR-RFL can dramatically improve the quality of reports for difficult samples. The source code will be available at https://github.com/SKD-HPC/LHR-RFL.
- Open Access
- 26 November 2024
- Published by Elsevier in Journal of Systems Architecture
- Vol. 158
ChipAI: A scalable chiplet-based accelerator for efficient DNN inference using silicon photonics
- 26 November 2024
- Published by Elsevier in Journal of Systems Architecture
- Vol. 158
Open AccessChipAI: A scalable chiplet-based accelerator for efficient DNN inference using silicon photonics
- Top Cited Papers
- 5 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 28 (4) , 2152-2162
TSGET: Two-Stage Global Enhanced Transformer for Automatic Radiology Report Generation
Recently, automatic radiology report generation, which targets to generate multiple sentences that can accurately describe medical observations for given X-ray images, has gained increasing attention. Existing methods commonly employ the attention mechanism for accurate word generation. However, such attention-based methods fail to leverage useful image-level global features, thereby limiting the model's reasoning ability. To tackle this challenge, we propose two-stage global enhancement layers to facilitate the Transformer to generate more reliable reports from a global perspective. Specifically, the $1^{st}$ Global Enhancement Layer ( $1^{st}$ GEL) is designed to capture the global visual context features by establishing the relationships between image-level global features and previously generated words. The $2^{nd}$ Global Enhancement Layer ( $2^{nd}$ GEL) is devised to capture the region-global level features by building the relationships between image-level global features and region-level information. The experiments demonstrate that by integrating the aforementioned two-stage global enhancement layers into the Transformer model, our proposal achieves state-of-the-art (SOTA) performance on various Natural Language Generation (NLG) evaluation metrics. Further Clinical Efficacy (CE) evaluations also validate that our proposal is able to predict more critical information. The Code will be available at https://github.com/SKD-HPC/TSGET .
- 5 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 28 (4) , 2152-2162
Top Cited PapersTSGET: Two-Stage Global Enhanced Transformer for Automatic Radiology Report Generation
Recently, automatic radiology report generation, which targets to generate multiple sentences that can accurately describe medical observations for given X-ray images, has gained increasing attention. Existing methods commonly employ the attention mechanism for accurate word generation. However, such attention-based methods fail to leverage useful image-level global features, thereby limiting the model's reasoning ability. To tackle this challenge, we propose two-stage global enhancement layers to facilitate the Transformer to generate more reliable reports from a global perspective. Specifically, the $1^{st}$ Global Enhancement Layer ( $1^{st}$ GEL) is designed to capture the global visual context features by establishing the relationships between image-level global features and previously generated words. The $2^{nd}$ Global Enhancement Layer ( $2^{nd}$ GEL) is devised to capture the region-global level features by building the relationships between image-level global features and region-level information. The experiments demonstrate that by integrating the aforementioned two-stage global enhancement layers into the Transformer model, our proposal achieves state-of-the-art (SOTA) performance on various Natural Language Generation (NLG) evaluation metrics. Further Clinical Efficacy (CE) evaluations also validate that our proposal is able to predict more critical information. The Code will be available at https://github.com/SKD-HPC/TSGET .
- 23 August 2023
- Published by Elsevier in Parallel Computing
- Vol. 117
ESA: An efficient sequence alignment algorithm for biological database search on Sunway TaihuLight
- 23 August 2023
- Published by Elsevier in Parallel Computing
- Vol. 117
ESA: An efficient sequence alignment algorithm for biological database search on Sunway TaihuLight
- Open Access
- 7 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 566-575
SEECHIP: A Scalable and Energy-Efficient Chiplet-based GPU Architecture Using Photonic Links
- 7 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 566-575
Open AccessSEECHIP: A Scalable and Energy-Efficient Chiplet-based GPU Architecture Using Photonic Links
- Open Access
- 1 January 2023
- Published by National Library of Serbia in Computer Science and Information Systems
- Vol. 20 (1) , 513-535
Efficient neural network accelerators with optical computing and communication
Conventional electronic Artificial Neural Networks (ANNs) accelerators focus on architecture design and numerical computation optimization to improve the training efficiency. However, these approaches have recently encountered bottlenecks in terms of energy efficiency and computing performance, which leads to an increase interest in photonic accelerator. Photonic architectures with low energy consumption, high transmission speed and high bandwidth have been considered as an important role for generation of computing architectures. In this paper, to provide a better understanding of optical technology used in ANN acceleration, we present a comprehensive review for the efficient photonic computing and communication in ANN accelerators. The related photonic devices are investigated in terms of the application in ANNs acceleration, and a classification of existing solutions is proposed that are categorized into optical computing acceleration and optical communication acceleration according to photonic effects and photonic architectures. Moreover, we discuss the challenges for these photonic neural network acceleration approaches to highlight the most promising future research opportunities in this field.
- 1 January 2023
- Published by National Library of Serbia in Computer Science and Information Systems
- Vol. 20 (1) , 513-535
Open AccessEfficient neural network accelerators with optical computing and communication
Conventional electronic Artificial Neural Networks (ANNs) accelerators focus on architecture design and numerical computation optimization to improve the training efficiency. However, these approaches have recently encountered bottlenecks in terms of energy efficiency and computing performance, which leads to an increase interest in photonic accelerator. Photonic architectures with low energy consumption, high transmission speed and high bandwidth have been considered as an important role for generation of computing architectures. In this paper, to provide a better understanding of optical technology used in ANN acceleration, we present a comprehensive review for the efficient photonic computing and communication in ANN accelerators. The related photonic devices are investigated in terms of the application in ANNs acceleration, and a classification of existing solutions is proposed that are categorized into optical computing acceleration and optical communication acceleration according to photonic effects and photonic architectures. Moreover, we discuss the challenges for these photonic neural network acceleration approaches to highlight the most promising future research opportunities in this field.
- 23 November 2022
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 79 (10) , 10725-10746
Comparing the performance of multi-layer perceptron training on electrical and optical network-on-chips
- 23 November 2022
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 79 (10) , 10725-10746
Comparing the performance of multi-layer perceptron training on electrical and optical network-on-chips
- Open Access
- 9 September 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 7, 128054-128063
Implementation of Hybrid Alignment Algorithm for Protein Database Search on the SW26010 Many-Core Processor
In biological research, biology sequence alignment algorithm aims to find similarities between sequences. As the size of biological database increases exponentially, the complexity of sequence alignment process also increases rapidly, which results in a large amount of computational time. The Sunway TaihuLight is the world’s first heterogeneous supercomputer with peak performance over 100 PFlops and provides a new hardware platform for database search. In this paper we present an efficient method of protein database search based on Sunway TaihuLight supercomputer. Furthermore, we also optimize protein database search on Sunway TaihuLight to give full play to the performance of the SW26010 processor. In our proposed approach, we design hybrid sequence alignment by combining the Smith-Waterman local alignment algorithm and the Needleman-Wunsch global alignment algorithm. The protein database search is paralleled by message passing interface (MPI) and accelerated thread library (Athread). Experiment results with the Swiss-Prot database show that our implementation can effectively leverage the SW26010 processor’s special hardware architecture and achieve a speedup to 15.91 times on a single node. In addition, we expand the scale to 64 nodes to test the scalability of the parallel method on the Sunway TaihuLight system, and the results show that our parallel implementation of protein database search have a good expansibility and reliability.
- 9 September 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 7, 128054-128063
Open AccessImplementation of Hybrid Alignment Algorithm for Protein Database Search on the SW26010 Many-Core Processor
In biological research, biology sequence alignment algorithm aims to find similarities between sequences. As the size of biological database increases exponentially, the complexity of sequence alignment process also increases rapidly, which results in a large amount of computational time. The Sunway TaihuLight is the world’s first heterogeneous supercomputer with peak performance over 100 PFlops and provides a new hardware platform for database search. In this paper we present an efficient method of protein database search based on Sunway TaihuLight supercomputer. Furthermore, we also optimize protein database search on Sunway TaihuLight to give full play to the performance of the SW26010 processor. In our proposed approach, we design hybrid sequence alignment by combining the Smith-Waterman local alignment algorithm and the Needleman-Wunsch global alignment algorithm. The protein database search is paralleled by message passing interface (MPI) and accelerated thread library (Athread). Experiment results with the Swiss-Prot database show that our implementation can effectively leverage the SW26010 processor’s special hardware architecture and achieve a speedup to 15.91 times on a single node. In addition, we expand the scale to 64 nodes to test the scalability of the parallel method on the Sunway TaihuLight system, and the results show that our parallel implementation of protein database search have a good expansibility and reliability.