- Machine Learning
Research in this area focuses on large-scale data analytics, deep learning modeling, and large-scale model training, which are critical to data-driven intelligence.
Research in this area focuses on techniques that help optimize complex systems and solve decision-making issues in the real world. For example, these techniques have produced remarkable results in user traffic-based recommendation and the allocation of inventories and resources.
- MindOpt Optimization SolverLearn more
The MindOpt optimization solver is designed to solve problems in operations and delivers general solver capabilities. MindOpt can find answers to issues such as linear programming, mixed integer programming, and non-linear programming. It also supports black-box optimization and online optimization. Solutions for mathematical modeling and application of solvers are also provided for different industries. The linear programming solver of MindOpt is available to the public and is ranked first in the Linear Programming Benchmark by Hans D. Mittelmann, a renowned scholar whose benchmarks for optimization software are widely accepted.
- "Daling" System for Computing Resource Optimization
The Daling system is an intelligent solution oriented towards computing infrastructure. Daling can be integrated into computing resource management systems to provide optimal computing resource usage plans by leveraging machine learning and optimization techniques. In this way, it delivers higher stability and utilization of computing resources.
The Daling system includes the following modules:
1）Intelligent O&M, such as predictive maintenance, and anomaly detection and alerting.
2）Application profiling, such as usage prediction and interference detection
3）Scheduling, such as optimal resource orchestration, online scheduling, batch scheduling, rescheduling, load balancing, and auto scaling
4）Resource planning, such as resource capacity planning, and infrastructure planning and simulation During 2017's Double 11, storage scheduling reduced the peak resource usage by 30% and reduced the overall resource usage by 25%. In addition, cluster scheduling raised the overall utilization of CPUs from 70% to 90%.
- "Longling" System for User Traffic-based Optimization in the Retail Industry
The Longling system analyzes user behavior to construct features. Paired with an online decision-making module that utilizes user traffic streams, Longling implements recommendations that are optimized and tailored to individual users. As such, smart decision-making based on user traffic can be achieved. Longling is applied to several retail-related services, such as Tmall (shopping), Youku (video streaming), Hema (retail), and Xianyu (secondhand-goods marketplace). During Double 11, Longling plays an integral role in providing personalized recommendations, with over 600 million calls per day.
Jingqiao Zhang earned his PhD at Rensselaer Polytechnic Institute. He served as the senior manager of an applied science team to develop personalized recommendation solutions for sales improvement and customer engagement of Amazon Devices such as Echo, Tablet, and Kindle. During his course of PhD, Jingqiao focused on evolutionary optimization and pattern recognition. He published a monograph and over 10 papers, with more than 1,000 citations. Jingqiao joined the Decision Intelligence Lab in 2018. He leads the development of self-supervised representation learning in big data scenarios.
Cheng Yang has over 10 years of experience in the fields of machine learning and data mining. In the past, he has overseen the development of a number of commercial algorithms, such as algorithms of search-based recommendation, intelligent marketing, and supply chain optimization. Cheng has been an active contributor to many large-scale real-time machine learning projects under extreme conditions such as Double 11. His latest research includes smart decision-making, such as online learning and dynamic optimization, in the new retail scenario.
Liang Sun earned his PhD in computer science at Arizona State University. He served Microsoft Azure Machine Learning Department as a senior data scientist. Liang carries out research in algorithms such as anomaly detection, drill-down attribution analysis, and resource optimization, as well as the application of these algorithms in intelligent maintenance and asset loss prevention. Liang has published close to 30 journal and conference papers in the fields of machine learning and data mining. He has also published an English monograph and a Chinese monograph both on machine learning.
- Qingsong Wen, Kai He, Liang Sun, Yingying Zhang, Min Ke, and Huan Xu, "RobustPeriod: Time-Frequency Mining for Robust Multiple Periodicities Detection," in Proc. ACM SIGMOD International Conference on Management of Data (SIGMOD 2021), Xi'an, China, Jun. 2021.
- Qingyang Xu, Qingsong Wen, Liang Sun, A Two-Stage Framework for Seasonal Time Series Forecasting, in Proc. of IEEE 46th International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2021), Toronto, Canada, June 2021
- Linxiao Yang, Qingsong Wen, Bo Yang, Liang Sun, A Robust and Efficient Multi-Scale Seasonal-Trend Decomposition, in Proc. of IEEE 46th International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2021), Toronto, Canada, June 2021.
- Qingsong Wen, Zhengzhi Ma, and Liang Sun, "On Robust Variance Filtering and Change Of Variance Detection," accepted for Oral Presentation, in Proc. of IEEE 45th International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2020), Barcelona, Spain, May 2020. Oral paper.
- Qingsong Wen, Zhe Zhang, Yan Li and Liang Sun, "Fast RobustSTL: Efficient and Robust Seasonal-Trend Decomposition for Time Series with Complex Patterns," in Proc. of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD 2020), San Diego, CA, Aug. 2020
- Yifei Zhao, Yu-Hang Zhou, Mingdong Ou, Huan Xu, Nan Li: Maximizing Cumulative User Engagement in Sequential Recommendation: An Online Optimization Perspective. KDD 2020: 2784-2792
- Guillermo Gallego, Anran Li, Van-Anh Truong, and Xinshang Wang. 2020. “Approximation Algorithms for Product Framing and Pricing.” Operations Research 68 (1): 134-60.
- David Simchi-Levi, Rui Sun, and Xinshang Wang. 2019. “Online Matching with Bayesian Rewards.” SSRN Electronic Journal.
- Rong Jin, David Simchi-Levi, Li Wang, Xinshang Wang, and Sen Yang. 2019. “Shrinking the Upper Confidence Bound: A Dynamic Product Selection Problem for Urban Warehouses.” SSRN Electronic Journal.
- Zhao Kui, Junhao Hua, Ling Yan, Qi Zhang, Huan Xu, and Cheng Yang. "A Unified Framework for Marketing Budget Allocation." In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pp. 1820-1830. 2019.
- Yi Peng, Miao Xie, Jiahao Liu, Xuying Meng, Nan Li, Cheng Yang, Tao Yao. A Practical Semi-Parametric Contextual Bandit. International Joint Conference on Artificial Intelligence 2019 (IJCAI)
- Mingdong Ou, Nan Li, Cheng Yang, Shenghuo Zhu, and Rong Jin. Semi-parametric sampling for stochasitc bandits with many arms. In Proc. of AAAI Conference on Artificial Intelligence 2019 (AAAI)
- Yu-Hang Zhou, Chen Liang, Nan Li, Cheng Yang, Shenghuo Zhu, Rong Jin. Robust Online Matching with User Arrival Distribution Drift. In Proc. of AAAI Conference on Artificial Intelligence 2019 (AAAI)
- Qingsong Wen, Jingkun Gao, Xiaomin Song, Liang Sun, Jian Tan. "RobustTrend: A Huber Loss with a Combined First and Second Order Difference Regularization for Time Series Trend Filtering," in Proc. of the 28th International Joint Conference on Artificial Intelligence (IJCAI 2019), pp. 3856-3862, Macao, China, Aug. 2019. Oral paper.
- Qingsong Wen, Jingkun Gao, Xiaomin Song, Liang Sun, Huan Xu, Shenghuo Zhu. "RobustSTL: A Robust Seasonal-Trend Decomposition Algorithm for Long Time Series," in Proc. of the 33rd AAAI Conference on Artificial Intelligence (AAAI 2019), 2019, pp. 5409-5416, Honolulu, Hawaii, Jan. 2019. Oral paper.
- Cong Leng, Hao Li, Shenghuo Zhu, Rong Jin. Extremely Low Bit Neural Network: Squeeze the Last Bit Out with ADMM. In: Proceedings of the 32rd AAAI Conference on Artificial Intelligence (AAAI, 18), New Orleans, LA, 2018.
- Ao Zhang, Nan Li, Jian Pu, Jun Wang, Junchi Yan, Hongyuan Zha. tau-FPL: Tolerance-Constrained Learning in Linear Time. In: Proceedings of the 32rd AAAI Conference on Artificial Intelligence (AAAI, 18), New Orleans, LA, 2018.
- Qi Qian, Jisheng Tang, Hao Li, Shenghuo Zhu and Rong Jin. Large-scale Distance Metric Learning with Uncertainty. In: Proceedings of the 31th IEEE Conference on Computer Vision and Pattern Recognition (CVPR, 18), Salt Lake City, UT, 2018.
- Mingdong Ou, Nan Li, Shenghuo Zhu, Rong Jin. Multinomial Logit Bandit with Linear Utility Functions. In: Proceedings of the 27th International Joint Conference on Artificial Intelligence (IJCAI, 18), 2018.
- Yang Yu, Wei-Yang Qu, Nan Li, and Zimin Guo. Open category classification by adversarial sample generation. In: Proceedings of the 26th International Joint Conference on Artificial Intelligence (IJCAI, 17), Melbourne, Australia, 2017.
- In June 2018, the Decision Intelligence Lab won first place in the PASCAL VOC Target Detection Competition, the world's leading visual algoritithm competition.