Yuanmiao Lin

Yuanmiao Lin

Auto-generated profile, based on publicly available information

Yuanmiao Lin

School of Integrated Circuits, Guangdong University of Technology, Guangzhou, China

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
24 February 2025 in IEEE Computer Architecture Letters
Verified
Journal Article
Closed access
YLYuanmiao Lin
Yuanmiao Lin
SFShansen Fu
Shansen Fu
XLXueming Li
Xueming Li
CYChaoming Yang
Chaoming Yang
RLRongfeng Li
Rongfeng Li
XXXiaoming Xiong
Xiaoming Xiong
Precision-scalable convolutional neural networks (CNNs) offer a promising solution to balance network accuracy and hardware efficiency, facilitating high-performance execution on embedded devices. However, the requirement for small fine-grained multiplication calculations in precision-scalable (PS) networks has resulted in limited exploration on FPGA platforms. It is found that the deployment of PS accelerators encounters the following challenges: LUT-based multiply-accumulates (MACs) fail to make full use of DSP, and DSP-based MACs support limited precision combinations and cannot efficiently utilize DSP. Therefore, this brief proposes a DSP-based precision-scalable MAC with hybrid dataflow that supports most precision combinations and ensures high-efficiency utilization of DSP and LUT resources. Evaluating on mixed 4b/8b VGG16, compared with 8b baseline, the proposed accelerator achieves 3.97× improvement in performance with only a 0.37% accuracy degradation. Additionally, compared with state-of-the-art accelerators, the proposed accelerator achieves 1.20×-2.69× improvement in DSP efficiency and 1.63×-6.34× improvement in LUT efficiency.

13 May 2025 in IEEE Transactions on Circuits and Systems II: Express Briefs
Verified
Journal Article
Closed access
YLYuanmiao Lin
Yuanmiao Lin
ZFZhaozhang Feng
Zhaozhang Feng
XLXueming Li
Xueming Li
HHHongmin Huang
Hongmin Huang
RZRuidian Zhan
Ruidian Zhan
XXXiaoming Xiong
Xiaoming Xiong
Bit-level sparsity (BLS) technique and mixed-precision (MP) quantization present promising solutions for achieving high-performance execution on edge devices. It is found that employing either BLS technique or MP quantization independently for weight compression yields only limited compression. This brief presents a novel perspective on combining BLS and MP quantization to compress weights by utilizing column-based compression with sign-magnitude representation (SMR). Furthermore, by employing bit-serial computation, we design a column-based multiply-accumulate (CBMAC) unit that supports both BLS calculations and MP computations. Experimental results demonstrate that on VGG16 and ResNet50 by combining BLS and MP quantization, we compress weights down to 1.78b and 1.96b with 0.1% and 0.9% accuracy loss compared with INT8 networks, and achieve over 4.49× and 4.08× reduction in model size. Furthermore, when deployed on the XC7Z100 FPGA board, we achieve 670.5 GOPS and 4.09× improvement in performance compared with INT8 non-sparsity VGG16. Additionally, compared with state-of-the-art accelerators, on VGG16, we achieve 1.07×-5.33× and 1.69×-5.77× improvements in LUT efficiency and Equivalent LUT efficiency, respectively.

17 April 2026 in IEEE Transactions on Circuits and Systems I: Regular Papers
Verified
Journal Article
Closed access
YLYuanmiao Lin
Yuanmiao Lin
ZZZijian Zhang
Zijian Zhang
XLXueming Li
Xueming Li
HHHongmin Huang
Hongmin Huang
HMHeng Mai
Heng Mai
XXXiaoming Xiong
Xiaoming Xiong
Mixed-precision quantization and unstructured pruning have emerged as two effective compression techniques, demonstrating great potential in reducing model size and computational cost in the deployment of convolutional neural networks (CNNs). However, their joint deployment still faces two major challenges: 1) The former introduces heterogeneous bit-widths in the bit-level, while the latter results in irregular sparsity in the value-level; their fundamentally incompatible data representations and computation patterns require two distinct types of hardware overhead to process them separately, which severely limits hardware execution efficiency. 2) Jointly applying both techniques often leads to notable accuracy degradation. In this paper, we propose a novel compression perspective that reinterprets zero-values generated by unstructured pruning as multiple consecutive 0-bits. We further introduce column-based bit-level sparsity, which provides a unified representation for weights after mixed-precision quantization and unstructured pruning, requiring only a single type of hardware overhead. Based on these techniques, we develop a hybrid compression framework that jointly optimizes model size, accuracy, and hardware implementation. Our method achieves weight/activation precision of 2.13b/4.06b on VGG16, delivering 7.40 $\times $ compression and 2.74 $\times $ speedup with 0.92% accuracy loss compared to the 8b baseline. Compared to state-of-the-art accelerators, our design achieves 1.12 $\times $ -6.23 $\times $ and 1.31 $\times $ -6.60 $\times $ improvements in energy efficiency and LUT efficiency when deploying VGG16 and ResNet50.

12 January 2026 in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Verified
Journal Article
Closed access
YLYuanmiao Lin
Yuanmiao Lin
KLKelong Liang
Kelong Liang
PZPeng Zhang
Peng Zhang
ZZZijian Zhang
Zijian Zhang
XLXueming Li
Xueming Li
XXXiaoming Xiong
Xiaoming Xiong
Quantization and sparsity are key techniques for CNN compression and acceleration. Independent exploration of value-based sparsity (VBS) or bit-level sparsity (BLS) in 8b weights yield limited gains. We analyze weight distributions of VGG16 and ResNet50 after 4b quantization and find over 80% zero-values with strong near-structured sparsity. On VGG16, 98.11% of 16-weight groups contain at least 8 zero-values. Based on this analysis, we propose a two-stage sparse compression (TSSC) method that integrates VBS and BLS to efficiently compress weights. In stage 1, we apply post-quantization pruning to enforce structured sparsity with at least 8 zero-values per 16 weights and utilize value-based structured sparsity (VBSS) to compress these 8 zero-values per group. In stage 2, among the remaining weights, we apply column-based BLS (CBBLS) to further compress remaining weights. Finally, a two-stage sparse accelerator, which integrates the TSSC method and corresponding hardware optimization (hardware-efficient activation selection unit and bit-serial multiply-accumulate), is implemented to accelerate the CNN models. Experimental results demonstrate that our proposed TSSC method can reduce the weights of VGG16 and ResNet50 to 0.99b and 0.95b while maintaining an acceptable accuracy drop. When deployed on FPGA, our accelerator achieves an 8.08× weight compression and 7.13× performance improvement compared with 8b non-sparse VGG16 model. Compared with state-of-the-art accelerators, we achieve 1.95×-7.97×, 2.00×-32.8× and 1.13×-6.53× improvement in LUT efficiency, equivalent LUT efficiency and energy efficiency, respectively.

8 October 2025 in ACM Transactions on Embedded Computing Systems
Verified
Journal Article
Open access
XHXianghong Hu
Xianghong Hu
CYChaoming Yang
Chaoming Yang
XLXueming Li
Xueming Li
RLRongfeng Li
Rongfeng Li
YLYuanmiao Lin
Yuanmiao Lin
XXXiaoming Xiong
Xiaoming Xiong
Currently, there are two mainstream acceleration methods; one is mixed precision and the other is sparsity. Few accelerators support both mixed precision and sparsity, and most enable precision configurations across layers rather than within a single layer. Furthermore, most of accelerators adopt the traditional two’s complement (2C) data representation method, and we found that 2C brings many invalid “1” when representing signed data, which brings more resources overhead for mixed precision and many invalid operations for bit-level sparsity. Therefore, we propose a high-efficiency accelerator featuring a precision-scalable Sign-Magnitude Processing Element (SM-PE), which adopts a data representation method of SM and can flexibly support various precision calculations (2, 4, 8 bits) and bit-level sparsity. In addition, a dynamic quantization algorithm named DoReFaLike and a bit-level column sparsity (BLCS) technique are proposed to improve the efficiency of SM-PEs. Under the same accuracy constraint, the sparsity rate of the SM scheme is 3.5× higher than that of the 2C format. The accelerator has been synthesized on a 55nm CMOS ASIC platform. When scaled to 28nm, experimental results show that the energy efficiency of the proposed accelerator reaches 15.50, 25.37, 101.54 TOPS/W with 8-bit, 4-bit, and 2-bit input activations, respectively, and weights represented in sparse 8-bit precision, operating at 400 MHz. Compared to state-of-the-art accelerators, the proposed design achieves a performance improvement of 1.1× to 3.9×.

23 February 2026 in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
Verified
Journal Article
Closed access
YLYuanmiao Lin
Yuanmiao Lin
PZPeng Zhang
Peng Zhang
ZZZijian Zhang
Zijian Zhang
XLXueming Li
Xueming Li
HHHongmin Huang
Hongmin Huang
XXXiaoming Xiong
Xiaoming Xiong
N:M structured sparsity is key to convolutional neural network (CNN) compression and acceleration, but two challenges remain. From the algorithm perspective, prior works have mainly focused on 8-bit quantized models, where N:M sparsity yields limited hardware efficiency. From the hardware perspective, the cost differences across N:M sparsity have not been analyzed. To address these issues, we present a unified algorithm–hardware co-design framework for 4-bit CNN acceleration. We show that 4-bit quantization induces over 80% zero weights and strongly structured sparsity, with over 95% of weight groups satisfying 4:8, 8:16, or 16:32 patterns. We propose a pruning-after-quantization (PAQ) algorithm that enforces strict N:M sparsity with minimal accuracy loss. We also analyze the hardware overhead of activation fetch units (AFUs) under different N:M sparsity patterns (4:8, 8:16, 16:32), revealing that the 4:8 AFU reduces look-up table (LUT) cost by up to 66.7% compared to 16:32. Finally, we introduce a 4-bit LUT-based sign-magnitude multiplier (LBSMM) requiring only 11 LUT6 resources, outperforming existing multipliers. Integrated on a Xilinx VCU118 field-programmable gate array (FPGA), our accelerator achieves $2.51\times $ – $12.89\times $ improvements in equivalent LUT efficiency over SOTA designs. The implementations of the PAQ algorithm and the RTL of LBSMM are available at https://github.com/haden-01/PAQ-and-LBSMM.git