Yanlong Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yanlong Zhang has a total of 25 co-authors. Here are the top co-authors by number of shared publications:
- Aidong Chen
College of Robotics, Beijing Union University, Beijing, China
3 shared publications - Jing Zhou
School of Cyberspace Security, Northwestern Polytechnical University, Xi’an 710072, China
3 shared publications - Shuo Wang
Beijing Microelectronics Technology Institute, Beijing 100076, China
3 shared publications - Jian Liu
Chinese Academy of Sciences, State Key Laboratory of Semiconductor Physics and Chip Technologies, Institute of Semiconductors, Beijing, China
2 shared publications - Weifeng Liu
Artificial Intelligence and High-Speed Circuits Laboratory, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, China
2 shared publications - Wenchang Li
Laboratory of Solid-State Optoelectronic Information Technology, Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100083, China
2 shared publications
Latest Publications
- Open Access
Adaptive Temporal-Segment Inference and Statistical Test-Time Adaptation for UAV RF Recognition
Unmanned aerial vehicle (UAV) control signals are often bursty or utilize frequency hopping, so the evidence available in a spectrogram varies over time. We study confidence-adaptive multi-segment early exit (CAMSE) and separately evaluate online statistical adaptation of batch normalization (OSA-BN). CAMSE sequentially processes temporal segments with a shared ResNet18 and stops when the top-two probability margin from averaged logits reaches a stage-specific threshold. Temporal-segment augmentation supports segment-wise inference. An eight-class RFUAV evaluation separates 82 source IQ files into training, tuning, and test partitions. At signal-to-noise ratios (SNRs) from −30 to −22 dB, CAMSE achieves 48.53% accuracy with 2.47 segments per sample, compared with 49.07% for fixed-five fusion, reducing backbone evaluations by 50.51%. Regular-SNR accuracy is 99.52% with 1.06 segments. The whole-image Frozen-BN reference achieves 60.53% extreme-SNR accuracy. Uniform and decreasing thresholds offer different accuracy–computation operating points. In the separate whole-image experiment, OSA-BN raises extreme-SNR accuracy to 76.13% without labels, SNR estimates, or gradient updates. On eight native-noise IQ windows, CAMSE reduces single-thread CPU latency from 498.41 to 378.97 ms relative to fixed-five inference. These results support adaptive stopping within multi-segment processing and statistical adaptation for whole-image recognition.
Open AccessAdaptive Temporal-Segment Inference and Statistical Test-Time Adaptation for UAV RF Recognition
Unmanned aerial vehicle (UAV) control signals are often bursty or utilize frequency hopping, so the evidence available in a spectrogram varies over time. We study confidence-adaptive multi-segment early exit (CAMSE) and separately evaluate online statistical adaptation of batch normalization (OSA-BN). CAMSE sequentially processes temporal segments with a shared ResNet18 and stops when the top-two probability margin from averaged logits reaches a stage-specific threshold. Temporal-segment augmentation supports segment-wise inference. An eight-class RFUAV evaluation separates 82 source IQ files into training, tuning, and test partitions. At signal-to-noise ratios (SNRs) from −30 to −22 dB, CAMSE achieves 48.53% accuracy with 2.47 segments per sample, compared with 49.07% for fixed-five fusion, reducing backbone evaluations by 50.51%. Regular-SNR accuracy is 99.52% with 1.06 segments. The whole-image Frozen-BN reference achieves 60.53% extreme-SNR accuracy. Uniform and decreasing thresholds offer different accuracy–computation operating points. In the separate whole-image experiment, OSA-BN raises extreme-SNR accuracy to 76.13% without labels, SNR estimates, or gradient updates. On eight native-noise IQ windows, CAMSE reduces single-thread CPU latency from 498.41 to 378.97 ms relative to fixed-five inference. These results support adaptive stopping within multi-segment processing and statistical adaptation for whole-image recognition.
- Open Access
- 4 April 2026
- Published by MDPI AG in Mathematical and Computational Applications
- Vol. 31 (2) , 55
SCAFormer: Side-Channel Analysis Based on a Transformer with Focal Modulation
With the rapid development of Internet technology, information security has become increasingly important. Cryptographic analysis techniques, especially side-channel analysis (SCA), pose a significant threat to security systems. The latest SCA technology mainly utilizes the physical leakage signals generated during the operation of encryption devices, such as power consumption, temperature and electromagnetic radiation. These signals themselves carry the physical characteristics of the device, which are related to the encryption algorithm. Among them, the power consumption trace remains the main target of modern SCA research. However, such trajectories often bring about some analytical difficulties, such as the data sequence being too long, the feature points being distributed sparsely, and the internal relationships of the data being complex. These challenges hinder effective analysis. While Transformer architectures are good at capturing long-range dependencies in sequential data, their high computational complexity limits practical deployment. To address this, we propose replacing the self-attention (SA) module in Transformers with a focal modulation module. This modification significantly reduces computational complexity and reduces computational operations during feature extraction, enabling efficient and accurate side-channel attacks. Experimental results on benchmark datasets (ASCAD, AES_RD, AES_HD, DPAv4) demonstrate the superiority of our approach. The proposed method achieves a reduction in training time compared to standard Transformer models, and achieves superior key recovery performance, outperforming existing state-of-the-art models.
- 4 April 2026
- Published by MDPI AG in Mathematical and Computational Applications
- Vol. 31 (2) , 55
Open AccessSCAFormer: Side-Channel Analysis Based on a Transformer with Focal Modulation
With the rapid development of Internet technology, information security has become increasingly important. Cryptographic analysis techniques, especially side-channel analysis (SCA), pose a significant threat to security systems. The latest SCA technology mainly utilizes the physical leakage signals generated during the operation of encryption devices, such as power consumption, temperature and electromagnetic radiation. These signals themselves carry the physical characteristics of the device, which are related to the encryption algorithm. Among them, the power consumption trace remains the main target of modern SCA research. However, such trajectories often bring about some analytical difficulties, such as the data sequence being too long, the feature points being distributed sparsely, and the internal relationships of the data being complex. These challenges hinder effective analysis. While Transformer architectures are good at capturing long-range dependencies in sequential data, their high computational complexity limits practical deployment. To address this, we propose replacing the self-attention (SA) module in Transformers with a focal modulation module. This modification significantly reduces computational complexity and reduces computational operations during feature extraction, enabling efficient and accurate side-channel attacks. Experimental results on benchmark datasets (ASCAD, AES_RD, AES_HD, DPAv4) demonstrate the superiority of our approach. The proposed method achieves a reduction in training time compared to standard Transformer models, and achieves superior key recovery performance, outperforming existing state-of-the-art models.
- Open Access
Full-Element Analysis of Side-Channel Leakage Dataset on Symmetric Cryptographic Advanced Encryption Standard
The application of deep learning in side-channel analysis faces critical challenges arising from dispersed public datasets—i.e., datasets collected from heterogeneous sources and platforms with varying formats, labeling schemes, and sampling settings—and insufficient sample distribution uniformity, characterized by imbalanced class distributions and long-tailed label samples. This paper presents a systematic analysis of symmetric cryptographic AES side-channel leakage datasets, examining how these issues impact the performance of deep learning-based side-channel analysis (DL-SCA) models. We analyze over 10 widely used datasets, including DPA Contest and ASCAD, and highlight key inconsistencies via visualization, statistical metrics, and model performance evaluations. For instance, the DPA_v4 dataset exhibits extreme label imbalance with a long-tailed distribution, while the ASCAD datasets demonstrate missing leakage features. Experiments conducted using CNN and Transformer models show that such imbalances lead to high accuracy for a few labels (e.g., label 14 in DPA_v4) but also extremely poor accuracy (<0.5%) for others, severely degrading generalization. We propose targeted improvements through enhanced data collection protocols, training strategies, and feature alignment techniques. Our findings emphasize that constructing balanced datasets covering the full key space is vital to achieving robust and generalizable DL-SCA performance. This work contributes both empirical insights and methodological guidance for standardizing the design of side-channel datasets.
Open AccessFull-Element Analysis of Side-Channel Leakage Dataset on Symmetric Cryptographic Advanced Encryption Standard
The application of deep learning in side-channel analysis faces critical challenges arising from dispersed public datasets—i.e., datasets collected from heterogeneous sources and platforms with varying formats, labeling schemes, and sampling settings—and insufficient sample distribution uniformity, characterized by imbalanced class distributions and long-tailed label samples. This paper presents a systematic analysis of symmetric cryptographic AES side-channel leakage datasets, examining how these issues impact the performance of deep learning-based side-channel analysis (DL-SCA) models. We analyze over 10 widely used datasets, including DPA Contest and ASCAD, and highlight key inconsistencies via visualization, statistical metrics, and model performance evaluations. For instance, the DPA_v4 dataset exhibits extreme label imbalance with a long-tailed distribution, while the ASCAD datasets demonstrate missing leakage features. Experiments conducted using CNN and Transformer models show that such imbalances lead to high accuracy for a few labels (e.g., label 14 in DPA_v4) but also extremely poor accuracy (<0.5%) for others, severely degrading generalization. We propose targeted improvements through enhanced data collection protocols, training strategies, and feature alignment techniques. Our findings emphasize that constructing balanced datasets covering the full key space is vital to achieving robust and generalizable DL-SCA performance. This work contributes both empirical insights and methodological guidance for standardizing the design of side-channel datasets.
- Open Access
- 15 May 2025
- Published by MDPI AG in Electronics
- Vol. 14 (10) , 2006
Side-Channel Profiling Attack Based on CNNs’ Backbone Structure Variant
This paper focuses on sorting out two backbone network variants’ development routes for CNNs and transformers based on deep learning methods for side-channel analysis. Combining parametric and quantitative structural analyses, it is preferred that improvement research be carried out on the structure of the continuation CNN variants. Firstly, the comparative reproduction of Zaid’s Efficient CNN and Wouters’ Simplified Efficient CNN will be carried out; second, ResNet will be designed and implemented based on residual structure and the improved design of feature-coded CNNs; finally, the four methods in DPA_ contest v4.1, AES_RD, AES_HD, and ASCAD public datasets will be used for extensive analysis experiments to explore the effects of six preprocessing roles at three scales and further explore the impact of data enhancement by dataset noise, offset, and amplitude scaling. The experimental results show that ResNet based on the residual structure and the coded feature CNN proposed in this paper exhibit better performance advantages than the previous methods. Preferring a preprocessing method for each dataset can continue to reduce the mean rank; continuing to overlay data enhancement methods can make the model easier to converge and increase the model’s generalization ability. To advance the research in the branch of CNNs and transformer variants in the SCA field, the model methods obtained from the above experiments and the datasets obtained from the processing have been made publicly available on GitHub.
- 15 May 2025
- Published by MDPI AG in Electronics
- Vol. 14 (10) , 2006
Open AccessSide-Channel Profiling Attack Based on CNNs’ Backbone Structure Variant
This paper focuses on sorting out two backbone network variants’ development routes for CNNs and transformers based on deep learning methods for side-channel analysis. Combining parametric and quantitative structural analyses, it is preferred that improvement research be carried out on the structure of the continuation CNN variants. Firstly, the comparative reproduction of Zaid’s Efficient CNN and Wouters’ Simplified Efficient CNN will be carried out; second, ResNet will be designed and implemented based on residual structure and the improved design of feature-coded CNNs; finally, the four methods in DPA_ contest v4.1, AES_RD, AES_HD, and ASCAD public datasets will be used for extensive analysis experiments to explore the effects of six preprocessing roles at three scales and further explore the impact of data enhancement by dataset noise, offset, and amplitude scaling. The experimental results show that ResNet based on the residual structure and the coded feature CNN proposed in this paper exhibit better performance advantages than the previous methods. Preferring a preprocessing method for each dataset can continue to reduce the mean rank; continuing to overlay data enhancement methods can make the model easier to converge and increase the model’s generalization ability. To advance the research in the branch of CNNs and transformer variants in the SCA field, the model methods obtained from the above experiments and the datasets obtained from the processing have been made publicly available on GitHub.
- 1 January 2018
- Published by AIP Publishing in AIP Conference Proceedings
- Vol. 1955 (1) , 040084
A dynamically reconfigurable multi-functional PLL for SRAM-based FPGA in 65nm CMOS technology
Phase-locked loops (PLL) have been widely utilized in FPGA as an important module for clock management. PLL with dynamic reconfiguration capability is always welcomed in FPGA design as it is able to decrease power consumption and simultaneously improve flexibility. In this paper, a multi-functional PLL with dynamic reconfiguration capability for 65nm SRAM-based FPGA is proposed. Firstly, configurable charge pump and loop filter are utilized to optimize the loop bandwidth. Secondly, the PLL incorporates a VCO with dual control voltages to accelerate the adjustment of oscillation frequency. Thirdly, three configurable dividers are presented for flexible frequency synthesis. Lastly, a configuration block with dynamic reconfiguration function is proposed. Simulation results demonstrate that the proposed multi-functional PLL can output clocks with configurable division ratio, phase shift and duty cycle. The PLL can also be dynamically reconfigured without affecting other parts’ running or halting the FPGA device.
- 1 January 2018
- Published by AIP Publishing in AIP Conference Proceedings
- Vol. 1955 (1) , 040084
A dynamically reconfigurable multi-functional PLL for SRAM-based FPGA in 65nm CMOS technology
Phase-locked loops (PLL) have been widely utilized in FPGA as an important module for clock management. PLL with dynamic reconfiguration capability is always welcomed in FPGA design as it is able to decrease power consumption and simultaneously improve flexibility. In this paper, a multi-functional PLL with dynamic reconfiguration capability for 65nm SRAM-based FPGA is proposed. Firstly, configurable charge pump and loop filter are utilized to optimize the loop bandwidth. Secondly, the PLL incorporates a VCO with dual control voltages to accelerate the adjustment of oscillation frequency. Thirdly, three configurable dividers are presented for flexible frequency synthesis. Lastly, a configuration block with dynamic reconfiguration function is proposed. Simulation results demonstrate that the proposed multi-functional PLL can output clocks with configurable division ratio, phase shift and duty cycle. The PLL can also be dynamically reconfigured without affecting other parts’ running or halting the FPGA device.