Lin Ding
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
Lin Ding has a total of 25 co-authors. Here are the top co-authors by number of shared publications:
- Jiliang Zhang
Nanjing University of Posts and Telecommunications, Nanjing, Jiangsu, China
6 shared publications - Penggao He
College of Computer Science and Electronic Engineering, Hunan University, Changsha, Hunan, CHINA
3 shared publications Weihong Xu
School of Computer and Communication Engineering and Hunan Provincial Key Laboratory of Intelligent Processing of Big Data on Transportation
2 shared publications- Yuantao Chen
School of Computer Science and Engineering, Hunan University of Information Technology, Changsha, China
2 shared publications - Baosheng Zhang
Xi'an Jiaotong University, Xi'an, China
1 shared publication - Chongyan Gu
Queen’s University Belfast
1 shared publication
Latest Publications
- Open Access
- 14 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 2476-2490
ROSETTA: Efficient and Accurate Privacy-Preserving LLM Decoding via Hybrid CKKS/TFHE Evaluation
- 14 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 2476-2490
Open AccessROSETTA: Efficient and Accurate Privacy-Preserving LLM Decoding via Hybrid CKKS/TFHE Evaluation
- Open Access
- 26 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-7
PIMA-SecDB: Processing-in-memory Acceleration for Fully Homomorphic Encryption Databases
Databases utilizing fully homomorphic encryption (FHE) are the leading approach to preserving privacy for outsourced data. However, each query necessitates a full scan of all encrypted entries, imposing severe limitations on data access and processing efficiency. To address this bottleneck, we propose PIMA-SecDB, a processing-in-memory (PIM) architecture for FHE-based databases designed through rigorous hardware-software co-design. PIMA-SecDB features a three-tier hierarchical structure: the DIMM-level aggregation module, the x8 chip level with in-memory functional units, and the bank group level for data storage—enabling high intra-module bandwidth and parallelism for efficient in-situ processing. Moreover, we have developed a set of data management mechanisms to adapt to the parallel computing capabilities of PIMA-SecDB and a quantitative approach for selecting PIM-compatible encryption parameters to reduce computational load and data footprint. The experiment shows that PIMA-SecDB offers performance enhancements from 2.14 to 7.89× compared to existing FHE accelerators.
- 26 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-7
Open AccessPIMA-SecDB: Processing-in-memory Acceleration for Fully Homomorphic Encryption Databases
Databases utilizing fully homomorphic encryption (FHE) are the leading approach to preserving privacy for outsourced data. However, each query necessitates a full scan of all encrypted entries, imposing severe limitations on data access and processing efficiency. To address this bottleneck, we propose PIMA-SecDB, a processing-in-memory (PIM) architecture for FHE-based databases designed through rigorous hardware-software co-design. PIMA-SecDB features a three-tier hierarchical structure: the DIMM-level aggregation module, the x8 chip level with in-memory functional units, and the bank group level for data storage—enabling high intra-module bandwidth and parallelism for efficient in-situ processing. Moreover, we have developed a set of data management mechanisms to adapt to the parallel computing capabilities of PIMA-SecDB and a quantitative approach for selecting PIM-compatible encryption parameters to reduce computational load and data footprint. The experiment shows that PIMA-SecDB offers performance enhancements from 2.14 to 7.89× compared to existing FHE accelerators.
- 9 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computers
- Vol. 75 (6) , 2128-2140
APACHE: A Processing-Near-Memory Architecture for Multi-Scheme Fully Homomorphic Encryption
Fully Homomorphic Encryption (FHE) allows one to outsource computation over encrypted data to untrusted servers without worrying about data breaching. Since FHE is known to be extremely computationally intensive, application-specific accelerators emerged as a powerful solution to narrow the performance gap. Nevertheless, due to the increasing complexities in FHE schemes per se and multi-scheme FHE algorithm designs in end-to-end privacy-preserving tasks, existing FHE accelerators often face the challenges of low hardware utilization rates and insufficient memory bandwidth. In this work, we present APACHE, a layered near-memory computing hierarchy tailored for multi-scheme FHE acceleration. By closely inspecting the data flow across different FHE schemes, we propose a layered near-memory computing architecture with fine-grained functional unit design to significantly enhance the utilization rates of both computational resources and memory bandwidth. In addition, we propose a multi-scheme operator compiler to efficiently schedule high-level FHE computations across lower-level functional units. In the experiment, we evaluated APACHE in various FHE applications, such as Lola MNIST, HELR, fully packed bootstrapping, and fully homomorphic processors. The results illustrate that APACHE outperforms state-of-the-art ASIC FHE accelerators by 10.63$\times$ to 35.47$\times$ over a variety of operator and application benchmarks.- 9 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computers
- Vol. 75 (6) , 2128-2140
APACHE: A Processing-Near-Memory Architecture for Multi-Scheme Fully Homomorphic Encryption
Fully Homomorphic Encryption (FHE) allows one to outsource computation over encrypted data to untrusted servers without worrying about data breaching. Since FHE is known to be extremely computationally intensive, application-specific accelerators emerged as a powerful solution to narrow the performance gap. Nevertheless, due to the increasing complexities in FHE schemes per se and multi-scheme FHE algorithm designs in end-to-end privacy-preserving tasks, existing FHE accelerators often face the challenges of low hardware utilization rates and insufficient memory bandwidth. In this work, we present APACHE, a layered near-memory computing hierarchy tailored for multi-scheme FHE acceleration. By closely inspecting the data flow across different FHE schemes, we propose a layered near-memory computing architecture with fine-grained functional unit design to significantly enhance the utilization rates of both computational resources and memory bandwidth. In addition, we propose a multi-scheme operator compiler to efficiently schedule high-level FHE computations across lower-level functional units. In the experiment, we evaluated APACHE in various FHE applications, such as Lola MNIST, HELR, fully packed bootstrapping, and fully homomorphic processors. The results illustrate that APACHE outperforms state-of-the-art ASIC FHE accelerators by 10.63$\times$ to 35.47$\times$ over a variety of operator and application benchmarks.- 8 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 44 (11) , 4100-4112
DA PUF for IoT Security with 0.02% BER at 0.96 to 1.44 V and –40 °C to 125 °C
Physical unclonable functions (PUFs) serve as lightweight hardware security primitives that leverage chip fabrication process variations for key generation and device authentication. However, the lack of reliable challenge-response pairs (CRPs) and unrestricted query capabilities in current PUF-enabled applications introduces security vulnerabilities in resource-constrained devices. This article introduces a novel dual-state analog PUF (DA PUF) design that is capable of operating as either a leakage PUF or an inverter PUF to represent the low- and high-state PUF, respectively. Our proposed bit configuration strategy (BCS) significantly improves the reliability of PUF, achieving a reliability rate exceeding 99.98% in various working environments and voltage conditions. Additionally, the aggregate of reliable bits for both states is 1.88 (for the high state) and 2.13 (for the low state) times greater than that of a singular state. Within a temperature range of -40 °C–125 °C, the temperature sensitivity is measured at 0. 000356%/°C (high state) and 0.000417%/°C (low state); and within a voltage range of 0.96–1.44 V, the voltage sensitivity is recorded at 0.05546%/V (high state) and 0.03342%/V (low state). In addition, we developed a lightweight and secure DA PUF-based authentication protocol that significantly increases the number of possible device certifications. Through the random oracle model, we demonstrate that our proposed protocol meets essential security and resistance requirements. Compared to recent relevant protocols, our approach offers lower computational and communication overhead, making it highly suitable for Internet of Things (IoT) devices.
- 8 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 44 (11) , 4100-4112
DA PUF for IoT Security with 0.02% BER at 0.96 to 1.44 V and –40 °C to 125 °C
Physical unclonable functions (PUFs) serve as lightweight hardware security primitives that leverage chip fabrication process variations for key generation and device authentication. However, the lack of reliable challenge-response pairs (CRPs) and unrestricted query capabilities in current PUF-enabled applications introduces security vulnerabilities in resource-constrained devices. This article introduces a novel dual-state analog PUF (DA PUF) design that is capable of operating as either a leakage PUF or an inverter PUF to represent the low- and high-state PUF, respectively. Our proposed bit configuration strategy (BCS) significantly improves the reliability of PUF, achieving a reliability rate exceeding 99.98% in various working environments and voltage conditions. Additionally, the aggregate of reliable bits for both states is 1.88 (for the high state) and 2.13 (for the low state) times greater than that of a singular state. Within a temperature range of -40 °C–125 °C, the temperature sensitivity is measured at 0. 000356%/°C (high state) and 0.000417%/°C (low state); and within a voltage range of 0.96–1.44 V, the voltage sensitivity is recorded at 0.05546%/V (high state) and 0.03342%/V (low state). In addition, we developed a lightweight and secure DA PUF-based authentication protocol that significantly increases the number of possible device certifications. Through the random oracle model, we demonstrate that our proposed protocol meets essential security and resistance requirements. Compared to recent relevant protocols, our approach offers lower computational and communication overhead, making it highly suitable for Internet of Things (IoT) devices.
- 24 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 72 (2) , 647-660
Meta: A Memory-Efficient Tri-Stage Polynomial Multiplication Accelerator Using 2D Coupled-BFUs
Polynomial multiplication (PM) is the computational bottleneck of lattice-based cryptography, such as post-quantum cryptography (PQC). Designing dedicated hardware accelerators for polynomial multiplication is an effective solution to improve the execution speed. However, current mainstream designs ignore the impact of computing array size, resulting in poor design flexibility and low memory utilization. To address these issues, we propose, a memory-efficient tri-stage PM accelerator. Our proposed tri-stage PM algorithm fuses all isolated substages into a unique stage named fused coefficient-wise multiplication (FCWM), ensuring efficient computation. Meanwhile, in different stages of the algorithm, the circuit of two-dimensional reconfigurable coupled butterfly units (2D-RCBFUs) is fine-grained reconfigured to improve resource utilization. Moreover, the low-complexity memory mapping scheme simplifies the address control logic and reduces the hardware overhead. can efficiently support the PM of an arbitrary power of two, which is impossible for previous designs using a 2D computing array. Compared with the state-of-the-art designs, our demonstrates the best memory utilization, achieving up to $10.0\times$ performance improvement.
- 24 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 72 (2) , 647-660
Meta: A Memory-Efficient Tri-Stage Polynomial Multiplication Accelerator Using 2D Coupled-BFUs
Polynomial multiplication (PM) is the computational bottleneck of lattice-based cryptography, such as post-quantum cryptography (PQC). Designing dedicated hardware accelerators for polynomial multiplication is an effective solution to improve the execution speed. However, current mainstream designs ignore the impact of computing array size, resulting in poor design flexibility and low memory utilization. To address these issues, we propose, a memory-efficient tri-stage PM accelerator. Our proposed tri-stage PM algorithm fuses all isolated substages into a unique stage named fused coefficient-wise multiplication (FCWM), ensuring efficient computation. Meanwhile, in different stages of the algorithm, the circuit of two-dimensional reconfigurable coupled butterfly units (2D-RCBFUs) is fine-grained reconfigured to improve resource utilization. Moreover, the low-complexity memory mapping scheme simplifies the address control logic and reduces the hardware overhead. can efficiently support the PM of an arbitrary power of two, which is impossible for previous designs using a 2D computing array. Compared with the state-of-the-art designs, our demonstrates the best memory utilization, achieving up to $10.0\times$ performance improvement.
- 12 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems II: Express Briefs
- Vol. 71 (4) , 2339-2343
SI PUF: An SRAM and Inverter-Based PUF With a Bit Error Rate of 0.0053% and 0.073/0.042 pJ/bit
The physical unclonable function (PUF) can generate a unique identifier for each chip, ideal for key generation and chip anti-counterfeiting. The reliability of PUF is paramount, and therefore is one of the significant challenges for PUF design. This paper proposes a novel SRAM-and Inverter-based PUF (SI PUF) that can operate as either an SRAM PUF or an inverter PUF, depending on the input configuration signal. A zero-overhead bit configuration strategy (BCS) is proposed to enhance the reliability of PUF. Moreover, the working voltage of the sub-threshold level and the well-designed discharge stage ensure that our SI PUF can operate with low power consumption. The tested results of chips fabricated in 40-nm CMOS show that our SI PUF has only 0.0053% of the worst bit error rate (BER) under working conditions of -50 to 125∘C and 0.75 to 1.5V, with 0.073/0.042 pJ/bit of power consumption. The low BER and energy overhead illustrate that our SI PUF is more suitable for resource-limited devices compared to DAC 2022 and JSSC 2020.
- 12 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems II: Express Briefs
- Vol. 71 (4) , 2339-2343
SI PUF: An SRAM and Inverter-Based PUF With a Bit Error Rate of 0.0053% and 0.073/0.042 pJ/bit
The physical unclonable function (PUF) can generate a unique identifier for each chip, ideal for key generation and chip anti-counterfeiting. The reliability of PUF is paramount, and therefore is one of the significant challenges for PUF design. This paper proposes a novel SRAM-and Inverter-based PUF (SI PUF) that can operate as either an SRAM PUF or an inverter PUF, depending on the input configuration signal. A zero-overhead bit configuration strategy (BCS) is proposed to enhance the reliability of PUF. Moreover, the working voltage of the sub-threshold level and the well-designed discharge stage ensure that our SI PUF can operate with low power consumption. The tested results of chips fabricated in 40-nm CMOS show that our SI PUF has only 0.0053% of the worst bit error rate (BER) under working conditions of -50 to 125∘C and 0.75 to 1.5V, with 0.073/0.042 pJ/bit of power consumption. The low BER and energy overhead illustrate that our SI PUF is more suitable for resource-limited devices compared to DAC 2022 and JSSC 2020.
- 21 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems II: Express Briefs
- Vol. 69 (11) , 4543-4547
A Lightweight and Machine-Learning-Resistant PUF Using Obfuscation-Feedback-Shift-Register
Physical Unclonable Function (PUF) is a lightweight hardware security primitive and suitable for device authentication in the Internet of Things. However, each strong PUF instance must store at least 106 reliable challenge-response pairs in the center nodes, which brings an excessive storage overhead since centers connect massive remote PUFs. In this brief, an obfuscation-feedback-shift-register (OFSR) PUF is designed, consisting of certain weak PUF cells working with an obfuscation mechanism. Meanwhile, it efficiently reduces the storage overhead and overcomes the collapse response caused by normal linear-feedback-shift-register, providing higher security. Experiments show OFSR PUF has ideal performance on reliability, uniqueness, uniformity, randomness, and good resistance to machine learning attacks.
- 21 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems II: Express Briefs
- Vol. 69 (11) , 4543-4547
A Lightweight and Machine-Learning-Resistant PUF Using Obfuscation-Feedback-Shift-Register
Physical Unclonable Function (PUF) is a lightweight hardware security primitive and suitable for device authentication in the Internet of Things. However, each strong PUF instance must store at least 106 reliable challenge-response pairs in the center nodes, which brings an excessive storage overhead since centers connect massive remote PUFs. In this brief, an obfuscation-feedback-shift-register (OFSR) PUF is designed, consisting of certain weak PUF cells working with an obfuscation mechanism. Meanwhile, it efficiently reduces the storage overhead and overcomes the collapse response caused by normal linear-feedback-shift-register, providing higher security. Experiments show OFSR PUF has ideal performance on reliability, uniqueness, uniformity, randomness, and good resistance to machine learning attacks.
- Open Access
- 16 October 2020
- Published by Wiley in Complexity
- Vol. 2020, 1-11
Improved Density Peaks Clustering Based on Natural Neighbor Expanded Group
Density peaks clustering (DPC) is an advanced clustering technique due to its multiple advantages of efficiently determining cluster centers, fewer arguments, no iterations, no border noise, etc. However, it does suffer from the following defects: (1) difficult to determine a suitable value of its crucial cutoff distance parameter, (2) the local density metric is too simple to find out the proper center(s) of the sparse cluster(s), and (3) it is not robust that parts of prominent density peaks are remotely assigned. This paper proposes improved density peaks clustering based on natural neighbor expanded group (DPC-NNEG). The cores of the proposed algorithm contain two parts: (1) define natural neighbor expanded (NNE) and natural neighbor expanded group (NNEG) and (2) divide all NNEGs into a goal number of sets as the final clustering result, according to the closeness degree of NNEGs. At the same time, the paper provides the measurement of the closeness degree. We compared the state of the art with our proposal in public datasets, including several complex and real datasets. Experiments show the effectiveness and robustness of the proposed algorithm.
- 16 October 2020
- Published by Wiley in Complexity
- Vol. 2020, 1-11
Open AccessImproved Density Peaks Clustering Based on Natural Neighbor Expanded Group
Density peaks clustering (DPC) is an advanced clustering technique due to its multiple advantages of efficiently determining cluster centers, fewer arguments, no iterations, no border noise, etc. However, it does suffer from the following defects: (1) difficult to determine a suitable value of its crucial cutoff distance parameter, (2) the local density metric is too simple to find out the proper center(s) of the sparse cluster(s), and (3) it is not robust that parts of prominent density peaks are remotely assigned. This paper proposes improved density peaks clustering based on natural neighbor expanded group (DPC-NNEG). The cores of the proposed algorithm contain two parts: (1) define natural neighbor expanded (NNE) and natural neighbor expanded group (NNEG) and (2) divide all NNEGs into a goal number of sets as the final clustering result, according to the closeness degree of NNEGs. At the same time, the paper provides the measurement of the closeness degree. We compared the state of the art with our proposal in public datasets, including several complex and real datasets. Experiments show the effectiveness and robustness of the proposed algorithm.
- Open Access
- 18 July 2020
- Published by Wiley in Computational Intelligence and Neuroscience
- Vol. 2020, 1-15
Density Peaks Clustering by Zero-Pointed Samples of Regional Group Borders
Density peaks clustering algorithm (DPC) has attracted the attention of many scholars because of its multiple advantages, including efficiently determining cluster centers, a lower number of parameters, no iterations, and no border noise. However, DPC does not provide a reliable and specific selection method of threshold (cutoff distance) and an automatic selection strategy of cluster centers. In this paper, we propose density peaks clustering by zero-pointed samples (DPC-ZPSs) of regional group borders. DPC-ZPS finds the subclusters and the cluster borders by zero-pointed samples (ZPSs). And then, subclusters are merged into individuals by comparing the density of edge samples. By iteration of the merger, the suitable dc and cluster centers are ensured. Finally, we compared state-of-the-art methods with our proposal in public datasets. Experiments show that our algorithm automatically determines cutoff distance and centers accurately.
- 18 July 2020
- Published by Wiley in Computational Intelligence and Neuroscience
- Vol. 2020, 1-15
Open AccessDensity Peaks Clustering by Zero-Pointed Samples of Regional Group Borders
Density peaks clustering algorithm (DPC) has attracted the attention of many scholars because of its multiple advantages, including efficiently determining cluster centers, a lower number of parameters, no iterations, and no border noise. However, DPC does not provide a reliable and specific selection method of threshold (cutoff distance) and an automatic selection strategy of cluster centers. In this paper, we propose density peaks clustering by zero-pointed samples (DPC-ZPSs) of regional group borders. DPC-ZPS finds the subclusters and the cluster borders by zero-pointed samples (ZPSs). And then, subclusters are merged into individuals by comparing the density of edge samples. By iteration of the merger, the suitable dc and cluster centers are ensured. Finally, we compared state-of-the-art methods with our proposal in public datasets. Experiments show that our algorithm automatically determines cutoff distance and centers accurately.