
刁博宇 男 硕导 中国科学院计算技术研究所
电子邮件: diaoboyu2012@ict.ac.cn
通信地址: 北京市海淀区科学院南路6号
邮政编码: 100190
研究领域
高效机器学习系统、低功耗持续学习系统、嵌入式智能系统
招生信息
每年在计算所和国科大杭高院智能学院招收硕士研究生2-3名,联合培养博士研究生1名。
招生专业
081201-计算机系统结构
085404-计算机技术
085410-人工智能
教育和工作经历
教育背景
2012-09~2018-06 中国科学院计算技术研究所 工学博士
2008-09~2012-06 北京理工大学计算机学院 工学学士
工作简历
2021-10~现在, 中国科学院计算技术研究所, 高级工程师
2018-07~2021-09,中国科学院计算技术研究所, 助理研究员
社会兼职
2024-10-01~现在,中国计算机学会, 高级会员
2024-08-16~现在,中国计算机学会分布式计算与系统专业委员会, 执行委员
专利与奖励
专利成果
一种针对动态形状算子的编译优化方法,2024,专利号:CN119201130
一种深度学习模型执行时间的评估方法,2023,专利号:CN117422957
基于结构搜索和知识蒸馏的卷积神经网络模型压缩方法和装置,2023,专利号:CN117114053
降低神经网络推理所需内存占用量的重计算方法和系统,2023,专利号:CN117009093
一种基于贪心策略反向通道剪枝的图像处理方法及装置,2023,专利号:CN116992945
一种基于深度强化学习的集群任务调度方法及系统,2022,专利号:CN115904652
一种轻量级多任务视频流实时推理方法及系统,2022,专利号:CN115661712
一种自适应加速算子融合的方法及装置,2022,专利号:CN115659281
基于CPU-GPU异构并行的地表形变反演方法与系统,2022,专利号:CN114814843
一种边缘数据管理方法、介质、边缘服务器和系统, 2021, 专利号: CN111770152
奖励情况
中国指挥控制学会科技进步奖, 二等奖, 部委级, 2020
计算所优秀研究人员, 研究所(学校), 2019,2022
北京市优秀毕业生, 省级, 2018
出版信息
代表性论文
[IPDPS'25] Hangda Liu, Boyu Diao*, Yu Yang, Wenxin Chen, Xiaohui Peng, Yongjun Xu, Gensor: A Graph-based Construction Tensor Compilation Method for Deep Learning, 39th IEEE International Parallel & Distributed Processing Symposium, 2025 (CCF-B)
[ICASSP'25] Zijia An, Boyu Diao*, Libo Huang, Ruiqi Liu, Zhulin An, Yongjun Xu, IOR: Inversed Objects Replay for Incremental Object Detection, 50th IEEE International Conference on Acoustics, Speech, and Signal Processing, 2025 (CCF-B)
[CCGRID'25] Chen Yan, Boyu Diao*, Hangda Liu, Zhulin An, Yongjun Xu, A Nonlinear Hash-based Optimization Method for SpMV on GPUs, 25th IEEE international Symposium on Cluster, Cloud and Internet Computing, 2025 (CCF-C)
[CCGRID'25] Junzhou Xu, Boyu Diao*, Shaobo Zhao, Chengxiang Qi, Ruisheng Wang, Yongjun Xu, A Sensitivity-Driven Expert Allocation Method in LoRA-MoE for Efficient Fine-Tuning, 25th IEEE international Symposium on Cluster, Cloud and Internet Computing, 2025 (CCF-C)
[AAAI'25] Weilun Feng, Haotong Qin, Chuanguang Yang, Zhulin An, Libo Huang, Boyu Diao, Fei Wang, Renshuai Tao, Yongjun Xu, Michele Magno, MPQ-DM: Mixed Precision Quantization for Extremely Low Bit Diffusion Models, 39th Annual AAAI Conference on Artificial Intelligence,2025 (CCF-A)
[NeurIPS'24] Ruiqi Liu, Boyu Diao*, Libo Huang, Zijia An, Zhulin An, Yongjun Xu, Continual Learning in the Frequency Domain, 38th Annual Conference on Neural Information Processing Systems, 2024 (CCF-A)
[AAAI'24] Libo Huang, Yan Zeng, Chuanguang Yang, Zhulin An, Boyu Diao, Yongjun Xu, eTag: Class-Incremental Learning via Embedding Distillation and Task-Oriented Generation, 38th AAAI Conference on Artificial Intelligence, 2024. (CCF-A)
[CVPR'24] Chuanguang Yang, Zhulin An, Libo Huang, Junyu Bi, Han Yang, Boyu Diao, Yongjun Xu, CLIP-KD: A Comprehensive Study of Distilling CLIP Models, The Thirty-Fourth IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024 (CCF-A)
[ACM MM'24] Weilun Feng, Chuanguang Yang, Zhulin An, Libo Huang, Boyu Diao, Fei Wang, Yongjun Xu, Relational diffusion distillation for efficient image generation. 32nd ACM International Conference on Multimedia, 2024 (CCF-A)
[JPDC] Lingfei Dai, Luqi Gong, Zhulin An, Yogjun Xu, Boyu Diao*, Sketch-fusion: A gradient compression method with multi-layer fusion for communication-efficient distributed training, JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING, 2024 (CCF-B)
[CompJ] Hangda Liu, Boyu Diao*, Wenxin Chen, Yongjun Xu, A resource-aware workload scheduling method for unbalanced GEMMs on GPUs, The Computer Journal, 2024 (CCF-B)
[NPC'24] Wenxin Chen, Boyu Diao*, Hangda Liu, Ruisheng Wang, Yongjun Xu, DTuner: a Construction-based Optimization Method for Dynamic Tensor Operators Accelerating, 20th The IFIP International Conference on Network and Parallel Computing, 2024 (CCF-C)
[ICPR'24] Yu Yang, Boyu Diao*, Hangda Liu, Qiyun Chen, Qi Wang, Yongjun Xu, An Evolutionary Search-Based Operator Fusion Method with Binary Representation for Deep Learning Inference Acceleration, 27th International Conference on Pattern Recognition, 2024 (CCF-C)
[ICPR'24] Qiyun Chen, Boyu Diao*, Yu Yang, Yongjun Xu, SCP: A Structure Combination Pruning method via Structured Sparse for Deep Convolutional Neural Networks, 27th International Conference on Pattern Recognition, 2024 (CCF-C)
[CompJ] Dong Dong, Hongxu Jiang*, Boyu Diao*, AKGF: Automatic Kernel Generation for DNN on CPU-FPGA, The Computer Journal, 2023 (CCF-B)
[Innovation] Qi Wang, Tingting Li, Fei Wang, Boyu Diao, Lei Zheng, Jincai Huang, How to prevent malicious use of intelligent unmanned swarms?, THE INNOVATION, 2023 (IF:33.2)
[ICCV'23] Kelu Yao, Jin Wang, Boyu Diao, Chao Li, Towards Understanding the Generalization of Deepfake Detectors from a Game-Theoretical View, IEEE/CVF International Conference on Computer Vision, 2023 (CCF-A)
[IJCNN'22] Jianrong Xu, Boyu Diao*, Bifeng Cui, Kang Yang, Chao Li, Yongjun Xu, PFGDF: Pruning Filter via Gaussian Distribution Feature for Deep Neural Networks Acceleration, International Joint Conference on Neural Networks, 2022(CCF-C)
[AAAI'22] Chao Li, Kelu Yao, Jin Wang, Boyu Diao, Yongjun Xu, Quanshi Zhang, Interpretable Generative Adversarial Networks, The 36th AAAI Conference on Artificial Intelligence, 2022 (CCF-A)
科研活动
科研项目
宽域飞行跨学科寻优加速技术,负责人,中国科学院计划,2024-07~2029-06
嵌入式高性能可组合计算技术, 负责人, 国家任务, 2022-06~2025-07
嵌入式智能计算系统加速与优化,负责人,地方任务,2021-06~2022-12
智能迁移与部署技术, 参与, 国家任务, 2021-12~2025-12
端侧实时分析方法研究, 负责人, 国家任务, 2019-02~2022-01
指导学生
现指导学生
闫晨 硕士研究生 081200-计算机科学与技术
张熙 硕士研究生 085400-电子信息