Xuan Shen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Xuan Shen has a total of 54 co-authors. Here are the top co-authors by number of shared publications:
- Yanzhi Wang
Northeastern University
10 shared publications - Geng Yuan
University of Georgia
6 shared publications - Zhenglun Kong
Department of Electrical and Computer Engineering, Northeastern University, Boston, USA
5 shared publications - Hao Tang
Peking University, Beijing, China
3 shared publications - Peiyan Dong
Department of Electrical and Computer Engineering, Northeastern University, Boston, MA, USA
3 shared publications - Wei Niu
University of Georgia, Athens, GA, USA
3 shared publications
Latest Publications
- 12 September 2026
- Published by Springer Nature
- p. 76-92
ScAle: Attention Head Scaling as a Minimal Adapter for Spatial Reasoning in Vision–Language Models
- 12 September 2026
- Published by Springer Nature
- p. 76-92
ScAle: Attention Head Scaling as a Minimal Adapter for Spatial Reasoning in Vision–Language Models
- Open Access
- 26 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 11-19
Efficient and Accurate Post-Training Sparsification of Large Language Models with Proximal Operators
- 26 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 11-19
Open AccessEfficient and Accurate Post-Training Sparsification of Large Language Models with Proximal Operators
- Open Access
- 8 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 189-204
Graph Convolutional Network Acceleration Using Adiabatic Superconductor Josephson Devices
- 8 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 189-204
Open AccessGraph Convolutional Network Acceleration Using Adiabatic Superconductor Josephson Devices
- 4 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 44 (4) , 1406-1419
TSLA: A Task-Specific Learning Adaptation for Semantic Segmentation on Autonomous Vehicles Platform
Autonomous driving platforms encounter diverse driving scenarios, each with varying hardware resources and precision requirements. Given the computational limitations of embedded devices, it is crucial to consider computing costs when deploying on target platforms like the DRIVE PX 2. Our objective is to customize the semantic segmentation network according to the computing power and specific scenarios of autonomous driving hardware. We implement dynamic adaptability through a three-tier control mechanism—width multiplier, classifier depth, and classifier kernel—allowing fine-grained control over model components based on hardware constraints and task requirements. This adaptability facilitates broad model scaling, targeted refinement of the final layers, and scenario-specific optimization of kernel sizes, leading to improved resource allocation and performance. Additionally, we leverage Bayesian Optimization with surrogate modeling to efficiently explore hyperparameter spaces under tight computational budgets. Our approach addresses scenario-specific and task-specific requirements through automatic parameter search, accommodating the unique computational complexity and accuracy needs of autonomous driving. It scales its multiply-accumulate operations (MACs) for task-specific learning adaptation (TSLA), resulting in alternative configurations tailored to diverse self-driving tasks. These TSLA customizations maximize computational capacity and model accuracy, optimizing hardware utilization.
- 4 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 44 (4) , 1406-1419
TSLA: A Task-Specific Learning Adaptation for Semantic Segmentation on Autonomous Vehicles Platform
Autonomous driving platforms encounter diverse driving scenarios, each with varying hardware resources and precision requirements. Given the computational limitations of embedded devices, it is crucial to consider computing costs when deploying on target platforms like the DRIVE PX 2. Our objective is to customize the semantic segmentation network according to the computing power and specific scenarios of autonomous driving hardware. We implement dynamic adaptability through a three-tier control mechanism—width multiplier, classifier depth, and classifier kernel—allowing fine-grained control over model components based on hardware constraints and task requirements. This adaptability facilitates broad model scaling, targeted refinement of the final layers, and scenario-specific optimization of kernel sizes, leading to improved resource allocation and performance. Additionally, we leverage Bayesian Optimization with surrogate modeling to efficiently explore hyperparameter spaces under tight computational budgets. Our approach addresses scenario-specific and task-specific requirements through automatic parameter search, accommodating the unique computational complexity and accuracy needs of autonomous driving. It scales its multiply-accumulate operations (MACs) for task-specific learning adaptation (TSLA), resulting in alternative configurations tailored to diverse self-driving tasks. These TSLA customizations maximize computational capacity and model accuracy, optimizing hardware utilization.
- 29 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. PP (02780070) , 1
HotaQ: Hardware Oriented Token Adaptive Quantization for Large Language Models
The Large Language Models (LLMs) have been popular and widely used in creative ways because of their powerful capabilities. However, the substantial model size and complexity prevent LLMs from being implemented on resourceconstrained computing devices efficiently. Recent works that utilize quantization to boost the on-device efficiency of LLMs show that 8-bit or lower weight quantization is feasible with minimal impact on task performance with full-precision activation. Nevertheless, weight-only quantization does not leverage the potential acceleration on general edge processors, which typically support 16x16 and 8x8 integer multipliers. Additionally, it cannot benefit from specialized devices like Field-Programmable Gate Arrays (FPGAs), which offer reconfigure computing to design custom multipliers of arbitrary bit width. In this paper, we propose HotaQ, a Hardware-oriented token adaptive Quantization framework for general LLMs, and implement end-to-end accelerators on multiple devices. Considering the hardware profiling and activation analysis, we first introduce a basic activation quantization strategy to balance the trade-off of task performance and real inference speed. Then we integrate the activation-aware token pruning workflow to reduce the outliers and the adverse impact on attentivity. For mobile devices, we notably design the SIMD-based 4-bit multiplier and the efficient TRIP matrix multiplication for mobile devices. For FPGAs, we specially utilize the DSP packing techniques for 4/8-bit systolic-array-based multipliers and introduce an analytical model for achieving the best performance with limited on-chip resources. We apply our framework on different scales of LLMs, including LLaMA, OPT, and BLOOM with 4/8-bit for the activation and 4-bit for the weight. Experiments show that HotaQ achieves quantization on both model weights and activations while maintaining task performance comparable to existing weight-only quantization methods or even FP16 models. In the 4/8-bit scenario, HotaQ achieves an on-device speedup of up to 2.5× and 5.2× compared to its FP16 counterparts across edge devices and FPGAs respectively, marking a pioneering advancement in this domain.
- 29 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. PP (02780070) , 1
HotaQ: Hardware Oriented Token Adaptive Quantization for Large Language Models
The Large Language Models (LLMs) have been popular and widely used in creative ways because of their powerful capabilities. However, the substantial model size and complexity prevent LLMs from being implemented on resourceconstrained computing devices efficiently. Recent works that utilize quantization to boost the on-device efficiency of LLMs show that 8-bit or lower weight quantization is feasible with minimal impact on task performance with full-precision activation. Nevertheless, weight-only quantization does not leverage the potential acceleration on general edge processors, which typically support 16x16 and 8x8 integer multipliers. Additionally, it cannot benefit from specialized devices like Field-Programmable Gate Arrays (FPGAs), which offer reconfigure computing to design custom multipliers of arbitrary bit width. In this paper, we propose HotaQ, a Hardware-oriented token adaptive Quantization framework for general LLMs, and implement end-to-end accelerators on multiple devices. Considering the hardware profiling and activation analysis, we first introduce a basic activation quantization strategy to balance the trade-off of task performance and real inference speed. Then we integrate the activation-aware token pruning workflow to reduce the outliers and the adverse impact on attentivity. For mobile devices, we notably design the SIMD-based 4-bit multiplier and the efficient TRIP matrix multiplication for mobile devices. For FPGAs, we specially utilize the DSP packing techniques for 4/8-bit systolic-array-based multipliers and introduce an analytical model for achieving the best performance with limited on-chip resources. We apply our framework on different scales of LLMs, including LLaMA, OPT, and BLOOM with 4/8-bit for the activation and 4-bit for the weight. Experiments show that HotaQ achieves quantization on both model weights and activations while maintaining task performance comparable to existing weight-only quantization methods or even FP16 models. In the 4/8-bit scenario, HotaQ achieves an on-device speedup of up to 2.5× and 5.2× compared to its FP16 counterparts across edge devices and FPGAs respectively, marking a pioneering advancement in this domain.
- Open Access
- 27 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-9
AyE-Edge: Automated Deployment Space Search Empowering Accuracy yet Efficient Real-Time Object Detection on the Edge
- 27 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-9
Open AccessAyE-Edge: Automated Deployment Space Search Empowering Accuracy yet Efficient Real-Time Object Detection on the Edge
- Open Access
- 23 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-2
Late Breaking Result: AQFP-aware Binary Neural Network Architecture Search
Adiabatic Quantum-Flux-Parametron (AQFP) is a superconducting logic with extremely high energy efficiency. Recent research has made initial strides toward developing AQFP accelerator. However several critical challenges from both the hardware and software side remain, preventing the design from being a comprehensive solution. This paper proposes an AQFP-aware binary neural network architecture search framework that leverages software-hardware co-optimization to eventually search the AQFP-adapted neural network and the corresponding hardware configuration, providing a feasible AQFP-based solution for binary neural network (BNN) acceleration. Experimental results show that our framework consistently outperforms the representative AQFP-based framework.
- 23 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-2
Open AccessLate Breaking Result: AQFP-aware Binary Neural Network Architecture Search
Adiabatic Quantum-Flux-Parametron (AQFP) is a superconducting logic with extremely high energy efficiency. Recent research has made initial strides toward developing AQFP accelerator. However several critical challenges from both the hardware and software side remain, preventing the design from being a comprehensive solution. This paper proposes an AQFP-aware binary neural network architecture search framework that leverages software-hardware co-optimization to eventually search the AQFP-adapted neural network and the corresponding hardware configuration, providing a feasible AQFP-based solution for binary neural network (BNN) acceleration. Experimental results show that our framework consistently outperforms the representative AQFP-based framework.
- Open Access
Peeling the Onion: Hierarchical Reduction of Data Redundancy for Efficient Vision Transformer Training
Vision transformers (ViTs) have recently obtained success in many applications, but their intensive computation and heavy memory usage at both training and inference time limit their generalization. Previous compression algorithms usually start from the pre-trained dense models and only focus on efficient inference, while time-consuming training is still unavoidable. In contrast, this paper points out that the million-scale training data is redundant, which is the fundamental reason for the tedious training. To address the issue, this paper aims to introduce sparsity into data and proposes an end-to-end efficient training framework from three sparse perspectives, dubbed Tri-Level E-ViT. Specifically, we leverage a hierarchical data redundancy reduction scheme, by exploring the sparsity under three levels: number of training examples in the dataset, number of patches (tokens) in each example, and number of connections between tokens that lie in attention weights. With extensive experiments, we demonstrate that our proposed technique can noticeably accelerate training for various ViT architectures while maintaining accuracy. Remarkably, under certain ratios, we are able to improve the ViT accuracy rather than compromising it. For example, we can achieve 15.2% speedup with 72.6% (+0.4) Top-1 accuracy on Deit-T, and 15.7% speedup with 79.9% (+0.1) Top-1 accuracy on Deit-S. This proves the existence of data redundancy in ViT.
Open AccessPeeling the Onion: Hierarchical Reduction of Data Redundancy for Efficient Vision Transformer Training
Vision transformers (ViTs) have recently obtained success in many applications, but their intensive computation and heavy memory usage at both training and inference time limit their generalization. Previous compression algorithms usually start from the pre-trained dense models and only focus on efficient inference, while time-consuming training is still unavoidable. In contrast, this paper points out that the million-scale training data is redundant, which is the fundamental reason for the tedious training. To address the issue, this paper aims to introduce sparsity into data and proposes an end-to-end efficient training framework from three sparse perspectives, dubbed Tri-Level E-ViT. Specifically, we leverage a hierarchical data redundancy reduction scheme, by exploring the sparsity under three levels: number of training examples in the dataset, number of patches (tokens) in each example, and number of connections between tokens that lie in attention weights. With extensive experiments, we demonstrate that our proposed technique can noticeably accelerate training for various ViT architectures while maintaining accuracy. Remarkably, under certain ratios, we are able to improve the ViT accuracy rather than compromising it. For example, we can achieve 15.2% speedup with 72.6% (+0.4) Top-1 accuracy on Deit-T, and 15.7% speedup with 79.9% (+0.1) Top-1 accuracy on Deit-S. This proves the existence of data redundancy in ViT.
- Top Cited Papers
- 3 November 2022
- Published by Springer Nature
- p. 620-640
SPViT: Enabling Faster Vision Transformers via Latency-Aware Soft Token Pruning
- 3 November 2022
- Published by Springer Nature
- p. 620-640
Top Cited PapersSPViT: Enabling Faster Vision Transformers via Latency-Aware Soft Token Pruning
- Open Access
- 1 July 2022
- Published by International Joint Conferences on Artificial Intelligence
- Vol. 6, 5928-5931
Real-Time Portrait Stylization on the Edge
In this work we demonstrate real-time portrait stylization, specifically, translating self-portrait into cartoon or anime style on mobile devices. We propose a latency-driven differentiable architecture search method, maintaining realistic generative quality. With our framework, we obtain 10× computation reduction on the generative model and achieve real-time video stylization on off-the-shelf smartphone using mobile GPUs.
- 1 July 2022
- Published by International Joint Conferences on Artificial Intelligence
- Vol. 6, 5928-5931
Open AccessReal-Time Portrait Stylization on the Edge
In this work we demonstrate real-time portrait stylization, specifically, translating self-portrait into cartoon or anime style on mobile devices. We propose a latency-driven differentiable architecture search method, maintaining realistic generative quality. With our framework, we obtain 10× computation reduction on the generative model and achieve real-time video stylization on off-the-shelf smartphone using mobile GPUs.