Hengbo LIU's
homepage
homepage
Senior Algorithm Engineer
Email: hengboliu@163.com
My name is Hengbo LIU, and I have been working as a Data Scientist at Siemens, focusing on digital transformation in the process industry and process optimization. Before that, I have worked at Alibaba's Damo Academy and Ant group, focuses on AI time series research and their applications in various domains, including Alibaba Cloud's native autoscaling, internal platforms like Taobao and Flink, financial Liquidity Management as well as the Energy Industry. Prior to this, I worked at ZTE Corporation in Nanjing, where I was involved in AIOps algorithm research for cloud services and the development of the GoldenDB financial database.
I hold a Master's degree in Engineering and a Bachelor's degree in Science from Nanjing University of Aeronautics and Astronautics. During my academic journey, I specialized in fault diagnosis and fault-tolerant algorithms, contributing to the service of China's high-speed CRH-series traction systems.
My work and research interests revolve around time series analysis, sequence models, anomaly detection, and their applications in AIOps, Business Intelligence, and the Energy Industry. I leverage machine learning, deep learning, data mining, and signal processing techniques to tackle various challenges in these fields.
Recently, My focus has been closely following LLMs (Large Language Models), exploring their applications in time series forecasting. I've been working on fine-tuning LoRA with frozen layers and patching with LLaMA to enhance prediction accuracy. Additionally, I'm learning to build practical application modules with LangChain for better promotion purposes.
I have had the privilege of presenting my research at renowned international conferences such as ICASSP, AAAI, and CIKM. Additionally, I have been involved in developing an electric power forecasting algorithm library. Our team was honored with the title of "Digital Economy Industrial Innovation Achievement" in the New Product Launch event at the 2022 Global Digital Economy Conference, recognizing our contributions and innovation in the field.
My Group: Alibaba Damo DI
Liu H, Ma Z, Yang L, et al. SADI: A Self-Adaptive Decomposed Interpretable Framework for Electric Load Forecasting Under Extreme Events[C]. ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023: 1-5.
Zhu, Z., Chen, W., Xia, R., Zhou, T., Niu, P., Peng, B., Wang, W., Liu, H., Ma, Z., Wen, Q., & Sun, L. (2023). eForecaster: Unifying Electricity Forecasting with Robust, Flexible, and Explainable Machine Learning Algorithms. Proceedings of the AAAI Conference on Artificial Intelligence, 37(13), 15630-15638. https://doi.org/10.1609/aaai.v37i13.26853
Zhang Y, Guan Z, Qian H, L Xu, H Liu, et al. CloudRCA: a root cause analysis framework for cloud computing platforms[C]. Proceedings of the 30th ACM International Conference on Information & Knowledge Management(CIKM). 2021: 4373-4382.
Liu H, Mao Z, Jiang B, et al. Robust fault-tolerant control design for induction motor with faults and disturbances[C]. 2016 35th Chinese Control Conference (CCC). IEEE, 2016: 6795-6800.
5."A universal Heterogeneous Forecasting System for Electricity Power Based on Automatic Maching Learning". 2023 (Patent under review)
6. 2020-2023, Four Alibaba's innovation proposals of Time Series Forecasting, Autoscaling and AIOps
7. 2022-2023 Time series Forecasting with Extremely Event published in Alibaba Tech blog