YCYuxuan Cai

Yuxuan Cai

*This is an algorithmically generated author record based on publicly available information.

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


Yuxuan Cai has a total of 18 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • SecFed: A Secure and Efficient Federated Learning Based on Multi-Key Homomorphic Encryption

    by Yuxuan Cai, Wenxiu Ding, Yuxuan Xiao, Zheng Yan, Ximeng Liu, Zhiguo Wan

    Federated Learning (FL) is widely used in various industries because it effectively addresses the predicament of isolated data island. However, eavesdroppers is capable of inferring user privacy from the gradients or models transmitted in FL. Homomorphic Encryption (HE) can be applied in FL to protect sensitive data owing to its computability over ciphertexts. However, traditional HE as a single-key system cannot prevent dishonest users from intercepting and decrypting the ciphertexts from cooperative users in FL. Guaranteeing privacy and efficiency in this multi-user scenario is still a challenging target. In this paper, we propose a secure and efficient Federated Learning scheme (SecFed) based on multi-key HE to preserve user privacy and delegate some operations to TEE to improve efficiency while ensuring security. Specifically, we design the first TEE-based multi-key HE cryptosystem (EMK-BFV) to support privacy-preserving FL and optimize operation efficiency. Furthermore, we provide an offline protection mechanism to ensure the normal operation of system with disconnected participants. Finally, we give their security proofs and show their efficiency and superiority through comprehensive simulations and comparisons with existing schemes. SecFed offers a 3x performance improvement over TEE-based scheme and a 2x performance improvement over HE-based solution.


  • Open Access

    Mobile or FPGA? A Comprehensive Evaluation on Energy Efficiency and a Unified Optimization Framework

    by Geng Yuan, Peiyan Dong, Mengshu Sun, Wei Niu, Zhengang Li, Yuxuan Cai, Yanyu Li, Jun Liu, Weiwen Jiang, Xue Lin, et al.

    Efficient deployment of Deep Neural Networks (DNNs) on edge devices (i.e., FPGAs and mobile platforms) is very challenging, especially under a recent witness of the increasing DNN model size and complexity. Model compression strategies, including weight quantization and pruning, are widely recognized as effective approaches to significantly reduce computation and memory intensities, and have been implemented in many DNNs on edge devices. However, most state-of-the-art works focus on ad hoc optimizations, and there lacks a thorough study to comprehensively reveal the potentials and constraints of different edge devices when considering different compression strategies. In this article, we qualitatively and quantitatively compare the energy efficiency of FPGA-based and mobile-based DNN executions using mobile GPU and provide a detailed analysis. Based on the observations obtained from the analysis, we propose a unified optimization framework using block-based pruning to reduce the weight storage and accelerate the inference speed on mobile devices and FPGAs, achieving high hardware performance and energy-efficiency gain while maintaining accuracy.


  • Open Access

    YOLObile: Real-Time Object Detection on Mobile Devices via Compression-Compilation Co-Design

    by Yuxuan Cai, Hongjia Li, Geng Yuan, Wei Niu, Yanyu Li, Xulong Tang, Bin Ren, Yanzhi Wang

    The rapid development and wide utilization of object detection techniques have aroused attention on both accuracy and speed of object detectors. However, the current state-of-the-art object detection works are either accuracy-oriented using a large model but leading to high latency or speed-oriented using a lightweight model but sacrificing accuracy. In this work, we propose YOLObile framework, a real-time object detection on mobile devices via compression-compilation co-design. A novel block-punched pruning scheme is proposed for any kernel size. To improve computational efficiency on mobile devices, a GPU-CPU collaborative scheme is adopted along with advanced compiler-assisted optimizations. Experimental results indicate that our pruning scheme achieves 14$\times$ compression rate of YOLOv4 with 49.0 mAP. Under our YOLObile framework, we achieve 17 FPS inference speed using GPU on Samsung Galaxy S20. By incorporating our proposed GPU-CPU collaborative scheme, the inference speed is increased to 19.1 FPS, and outperforms the original YOLOv4 by 5$\times$ speedup. Source code is at: \url{https://github.com/nightsnack/YOLObile}.