Kelix Technical Report.

Jin Ouyang

Boyang Ding Chenglong Chu Dunju Zang Han Li 0005 Jiangxia Cao Kun Gai Muhao Wei Ruiming Tang Shiyao Wang 0001 Siyang Mao Xinchen Luo Yahui Liu Zhixin Ling Zhuoran Yang Ziming Li Chengru Song Guorui Zhou Guowang Zhang Hao Peng Hao Wang 0003 Jiaxin Deng Jin Ouyang Jinghao Zhang Lejian Ren Qianqian Wang Qigen Hu Tao Wang 0014 Xing-wei Wang 0001 Yiping Yang Zixing Zhang 0008 Ziqi Wang Kelix Technical Report. 2026 February abs/2602.09843 CoRR https://doi.org/10.48550/arXiv.2602.09843 db/journals/corr/corr2602.html#abs-2602-09843 streams/journals/corr

Enxia Li Jin Ouyang Sheng Xiang 0001 Lu Qin 0001 Ling Chen 0006 Efficient relation-aware heterogeneous graph neural network for fraud detection. 55 2025 September 28 World Wide Web (WWW) 5 https://doi.org/10.1007/s11280-025-01369-5 db/journals/www/www28.html#LiOXQC25 streams/journals/www

Yuzheng Zhang Renyu Yang Junhong Liu Weihan Jiang Tianyu Ye Yiqiao Liao Penghao Zhang Tiezi Zhang Kun Shang Tianyu Wo Chunming Hu Chengru Song Jin Ouyang Cuckoo: Deadline-Aware Job Packing on Heterogeneous GPUs for DL Model Training. 847-859 2025 SoCC https://doi.org/10.1145/3772052.3772266 conf/cloud/2025 db/conf/cloud/socc2025.html#ZhangYLJYLZZSWH25 Yihui Zhang Han Shen Renyu Yang Di Tian Yuxi Luo Menghao Zhang 0001 Li Li 0029 Chunming Hu Tianyu Wo Chengru Song Jin Ouyang Cauchy: A Cost-Efficient LLM Serving System through Adaptive Heterogeneous Deployment. 881-893 2025 SoCC https://doi.org/10.1145/3772052.3772264 conf/cloud/2025 db/conf/cloud/socc2025.html#ZhangSYTL00HWSO25 Ming Gao Guowang Zhang Jin Ouyang Nong Xiao 0001 LLM4Rec-LIGHTNING: High-Throughput Training System for LLM4Rec on Memory-Constrained GPUs. ICA3PP (8) 2025 187-197 conf/ica3pp/2025-8 https://doi.org/10.1007/978-981-95-8420-8_15 db/conf/ica3pp/ica3pp2025-8.html#GaoZOX25 streams/conf/ica3pp Jin Ouyang Xu Jia Sheng Xiang 0001 Ying Zhang 0001 Lu Qin 0001 Hypergraph-Driven Tabular Data Synthesis with Multi-objective Optimization. 451-467 2025 ICDAR (5) https://doi.org/10.1007/978-3-032-04630-7_26 conf/icdar/2025-5 db/conf/icdar/icdar2025-5.html#OuyangJXZQ25 streams/conf/icdar Zeying Wang Junhong Liu Penghao Zhang Xiaoyang Sun Xu Wang 0007 Tianyu Wo Chunming Hu Chengru Song Jin Ouyang Renyu Yang KAIOPS: A Platform Solution of End-to-End Multi-Modal AIOps for AI Training at Scale. 3192-3203 2025 ASE https://doi.org/10.1109/ASE63991.2025.00263 conf/kbse/2025 db/conf/kbse/ase2025.html#WangLZSWWHSOY25 streams/conf/kbse Yitang Yang Junhong Liu Jiapeng Chen Xiaoyang Sun Tianyu Wo Chunming Hu Chengru Song Jin Ouyang Renyu Yang Kair: A Statistical and Causal Approach to Pinpointing Stragglers in Distributed Model Training. 3754-3759 2025 ASE https://doi.org/10.1109/ASE63991.2025.00311 conf/kbse/2025 db/conf/kbse/ase2025.html#YangLCSWHSOY25 streams/conf/kbse

Kwai Keye Team Biao Yang Bin Wen Changyi Liu Chenglong Chu Chengru Song Chongling Rao Chuan Yi Da Li 0003 Dunju Zang Fan Yang 0094 Guorui Zhou Hao Peng Haojie Ding Jiaming Huang Jiangxia Cao Jiankang Chen Jingyun Hua Jin Ouyang Kaibing Chen Kaiyu Jiang Kaiyu Tang Kun Gai Shengnan Zhang Siyang Mao Sui Huang Tianke Zhang Tingting Gao Wei Chen 0070 Wei Yuan Xiangyu Wu Xiao Hu Xingyu Lu Yang Zhou Yifan Zhang 0004 Yiping Yang Yulong Chen Zhenhua Wu Zhenyu Li Zhixin Ling Ziming Li Dehua Ma Di Xu Haixuan Gao Hang Li 0007 Jiawei Guo Jing Wang Lejian Ren Muhao Wei Qianqian Wang Qigen Hu Shiyao Wang 0001 Tao Yu Xinchen Luo Yan Li 0043 Yiming Liang Yuhang Hu Zeyi Lu Zhuoran Yang Zixing Zhang 0008 Kwai Keye-VL Technical Report. 2025 July abs/2507.01949 CoRR https://doi.org/10.48550/arXiv.2507.01949 db/journals/corr/corr2507.html#abs-2507-01949 streams/journals/corr

Enxia Li Jin Ouyang Sheng Xiang 0001 Lu Qin 0001 Ling Chen 0006 Relation-Aware Heterogeneous Graph Neural Network for Fraud Detection. 240-255 2024 APWeb/WAIM (3) https://doi.org/10.1007/978-981-97-7238-4_16 conf/apweb/2024-3 db/conf/apweb/apweb2024-3.html#LiOXQC24 streams/conf/apweb Ziyang Liu Renyu Yang Jin Ouyang Weihan Jiang Tianyu Ye Menghao Zhang 0001 Sui Huang Jiaming Huang Chengru Song Di Zhang 0026 Tianyu Wo Chunming Hu Kale: Elastic GPU Scheduling for Online DL Model Training. 36-51 2024 SoCC https://doi.org/10.1145/3698038.3698532 conf/cloud/2024 db/conf/cloud/socc2024.html#LiuYOJY0HHSZWH24 Jianyong Zhu Renyu Yang Chunming Hu Tianyu Wo Shiqing Xue Jin Ouyang Jie Xu 0007 Perph: A Workload Co-location Agent with Online Performance Prediction and Resource Inference. 176-185 2021 CCGRID https://doi.org/10.1109/CCGrid51090.2021.00027 conf/ccgrid/2021 db/conf/ccgrid/ccgrid2021.html#ZhuYHWXOX21 Jianyong Zhu Renyu Yang Chunming Hu Tianyu Wo Shiqing Xue Jin Ouyang Jie Xu 0007 Perphon: a ML-based Agent for Workload Co-location via Performance Prediction and Resource Inference. 478 2019 SoCC https://doi.org/10.1145/3357223.3365440 https://www.wikidata.org/entity/Q131119374 conf/cloud/2019 db/conf/cloud/socc2019.html#ZhuYHWXO019

Renyu Yang Yang Zhang Peter Garraghan Yihui Feng Jin Ouyang Jie Xu 0007 Zhuo Zhang 0015 Chao Li Reliable Computing Service in Massive-Scale Systems through Rapid Low-Cost Failover. 969-983 2017 10 IEEE Trans. Serv. Comput. 6 https://doi.org/10.1109/TSC.2016.2544313 http://doi.ieeecomputersociety.org/10.1109/TSC.2016.2544313 https://www.wikidata.org/entity/Q59886671 db/journals/tsc/tsc10.html#YangZGFOXZL17

Jia Zhan Jin Ouyang Fen Ge Jishen Zhao Yuan Xie 0001 Hybrid Drowsy SRAM and STT-RAM Buffer Designs for Dark-Silicon-Aware NoC. 3041-3054 2016 24 IEEE Trans. Very Large Scale Integr. Syst. 10 https://doi.org/10.1109/TVLSI.2016.2536747 db/journals/tvlsi/tvlsi24.html#ZhanOGZ016

Jia Zhan Jin Ouyang Fen Ge Jishen Zhao Yuan Xie 0001 DimNoC: a dim silicon approach towards power-efficient on-chip network. 10:1-10:6 2015 DAC https://doi.org/10.1145/2744769.2744824 conf/dac/2015 db/conf/dac/dac2015.html#ZhanOGZ015

Jia Zhan Nikolay Stoimenov Jin Ouyang Lothar Thiele Vijaykrishnan Narayanan Yuan Xie 0001 Optimizing the NoC Slack Through Voltage and Frequency Scaling in Hard Real-Time Embedded Systems. 1632-1643 2014 33 IEEE Trans. Comput. Aided Des. Integr. Circuits Syst. 11 https://doi.org/10.1109/TCAD.2014.2347921 db/journals/tcad/tcad33.html#ZhanSOTN014

Jin Ouyang Thomas Weise 0001 Alexandre Devert Raymond Chiong SDGP: A developmental approach for traveling salesman problems. 78-85 2013 CIPLS https://doi.org/10.1109/CIPLS.2013.6595203 conf/cipls/2013 db/conf/cipls/cipls2013.html#OuyangWDC13 Jia Zhan Nikolay Stoimenov Jin Ouyang Lothar Thiele Vijaykrishnan Narayanan Yuan Xie 0001 Designing energy-efficient NoC for real-time embedded systems through slack optimization. 37:1-37:6 2013 DAC https://doi.org/10.1145/2463209.2488780 conf/dac/2013 db/conf/dac/dac2013.html#ZhanSOTNX13

Guangyu Sun 0003 Yibo Chen Xiangyu Dong 0001 Jin Ouyang Yuan Xie 0001 Three-dimensional Integrated Circuits: Design, EDA, and Architecture. 1-151 2011 5 Found. Trends Electron. Des. Autom. 1-2 https://doi.org/10.1561/1000000016 db/journals/fteda/fteda5.html#SunCDOX11

Guangyu Sun 0003 Dimin Niu Jin Ouyang Yuan Xie 0001 A frequent-value based PRAM memory architecture. 211-216 2011 ASP-DAC https://doi.org/10.1109/ASPDAC.2011.5722186 http://dl.acm.org/citation.cfm?id=1950867 conf/aspdac/2011 db/conf/aspdac/aspdac2011.html#SunNOX11 Jin Ouyang Yuan Xie 0001 Enabling quality-of-service in nanophotonic network-on-chip. 351-356 2011 ASP-DAC https://doi.org/10.1109/ASPDAC.2011.5722212 http://dl.acm.org/citation.cfm?id=1950891 conf/aspdac/2011 db/conf/aspdac/aspdac2011.html#OuyangX11 Jin Ouyang Chuan Yang Dimin Niu Yuan Xie 0001 Zhiwen Liu F<sup>2</sup>BFLY: an on-chip free-space optical network with wavelength-switching. 348-358 2011 ICS https://doi.org/10.1145/1995896.1995949 conf/ics/2011 db/conf/ics/ics2011.html#OuyangYNXL11 Jin Ouyang Jing Xie 0006 Matthew Poremba Yuan Xie 0001 Evaluation of using inductive/capacitive-coupling vertical interconnects in 3D network-on-chip. 477-482 2010 ICCAD https://doi.org/10.1109/ICCAD.2010.5653769 https://doi.ieeecomputersociety.org/10.1109/ICCAD.2010.5653769 http://dl.acm.org/citation.cfm?id=2133529 conf/iccad/2010 db/conf/iccad/iccad2010.html#OuyangXPX10 Jin Ouyang Yuan Xie 0001 LOFT: A High Performance Network-on-Chip Providing Quality-of-Service Support. 409-420 2010 MICRO https://doi.org/10.1109/MICRO.2010.21 https://doi.ieeecomputersociety.org/10.1109/MICRO.2010.21 http://dl.acm.org/citation.cfm?id=1934986 conf/micro/2010 db/conf/micro/micro2010.html#OuyangX10 Jin Ouyang Guangyu Sun 0003 Yibo Chen Lian Duan Tao Zhang 0032 Yuan Xie 0001 Mary Jane Irwin Arithmetic unit design using 180nm TSV-based 3D stacking technology. 1-4 2009 3DIC https://doi.org/10.1109/3DIC.2009.5306565 conf/3dic/2009 db/conf/3dic/3dic2009.html#OuyangSCDZXI09 Yaoyao Ye Lian Duan Jiang Xu 0001 Jin Ouyang Mo Kwai Hung Yuan Xie 0001 3D optical networks-on-chip (NoC) for multiprocessor systems-on-chip (MPSoC). 1-6 2009 3DIC https://doi.org/10.1109/3DIC.2009.5306588 conf/3dic/2009 db/conf/3dic/3dic2009.html#YeDXOHX09 Jin Ouyang Raghuveer Raghavendra Sibin Mohan Tao Zhang 0032 Yuan Xie 0001 Frank Mueller 0001 CheckerCore: enhancing an FPGA soft core to capture worst-case execution times. 175-184 2009 CASES https://doi.org/10.1145/1629395.1629421 conf/cases/2009 db/conf/cases/cases2009.html#OuyangRMZXM09 Yibo Chen Jin Ouyang Yuan Xie 0001 ILP-based scheme for timing variation-aware scheduling and resource binding. 27-30 2008 SoCC https://doi.org/10.1109/SOCC.2008.4641473 conf/socc/2008 db/conf/socc/socc2008.html#ChenOX08 Jin Ouyang Yuan Xie 0001 Power optimization for FinFET-based circuits using genetic algorithms. 211-214 2008 SoCC https://doi.org/10.1109/SOCC.2008.4641513 conf/socc/2008 db/conf/socc/socc2008.html#OuyangX08 Jiangxia Cao Jiankang Chen Jiapeng Chen Kaibing Chen Ling Chen 0006 Wei Chen 0070 Yibo Chen Yulong Chen Raymond Chiong Chenglong Chu Jiaxin Deng Alexandre Devert Boyang Ding Haojie Ding Xiangyu Dong 0001 Lian Duan Yihui Feng Kun Gai Haixuan Gao Ming Gao Tingting Gao Peter Garraghan Fen Ge Jiawei Guo Chunming Hu Qigen Hu Xiao Hu Yuhang Hu Jingyun Hua Jiaming Huang Sui Huang Mo Kwai Hung Mary Jane Irwin Kaiyu Jiang Weihan Jiang Chao Li Da Li 0003 Enxia Li Han Li 0005 Hang Li 0007 Li Li 0029 Yan Li 0043 Zhenyu Li Ziming Li Yiming Liang Yiqiao Liao Zhixin Ling Changyi Liu Junhong Liu Yahui Liu Zhiwen Liu Ziyang Liu Xingyu Lu Zeyi Lu Xinchen Luo Yuxi Luo Dehua Ma Siyang Mao Sibin Mohan Frank Mueller 0001 Dimin Niu Hao Peng Matthew Poremba Lu Qin 0001 Raghuveer Raghavendra Chongling Rao Lejian Ren Kun Shang Han Shen Chengru Song Nikolay Stoimenov Guangyu Sun 0003 Xiaoyang Sun Kaiyu Tang Ruiming Tang Kwai Keye Team Lothar Thiele Di Tian Narayanan VijaykrishnanVijaykrishnan Narayanan Hao Wang 0003 Jing Wang Qianqian Wang Shiyao Wang 0001 Tao Wang 0014 Xingwei Wang 0001Xing-wei Wang 0001 Xu Wang 0007 Zeying Wang Ziqi Wang Muhao Wei Thomas Weise 0001 Bin Wen Tianyu Wo Xiangyu Wu Zhenhua Wu Sheng Xiang 0001 Nong Xiao 0001 Jing Xie 0006 Yuan Xie 0001 Di Xu Jia XuXu Jia Jiang Xu 0001 Jie Xu 0007 Shiqing Xue Biao Yang Chuan Yang Fan Yang 0094 Renyu Yang Yiping Yang Yitang Yang Zhuoran Yang Tianyu Ye Yaoyao Ye Chuan Yi Tao Yu Wei Yuan Dunju Zang Jia Zhan Di Zhang 0026 Guowang Zhang Jinghao Zhang Menghao Zhang 0001 Penghao Zhang Shengnan Zhang Tao Zhang 0032 Tianke Zhang Tiezi Zhang Yang Zhang Yifan Zhang 0004 Yihui Zhang Ying Zhang 0001 Yuzheng Zhang Zhuo Zhang 0015 Zixing Zhang 0008 Jishen Zhao Guorui Zhou Yang Zhou Jianyong Zhu