CCF Ubiquitous Computing Committee Excellent Doctoral Forum
August 23 (Saturday) 15:30-17:30
Venue: 3rd Floor VIP Hall

introduction:
       Ning Jingyi, Ph.D. from Nanjing University under the supervision of Prof. Xie Lei, is an Assistant Researcher at the School of Computer Science, Nanjing University, as well as a Yuxiu Young Scholar and Zijin Scholar. Her research focuses on intelligent perception and mobile computing. In recent years, she has published over 20 papers in top international conferences/journals such as IEEE JSAC, IEEE TMC, ACM MobiCom, ACM UbiComp, and IEEE INFOCOM, including three first-author papers in ACM MobiCom. She has been a key contributor to major projects funded by the National Natural Science Foundation of China (NSFC), including key programs, special projects, and Jiangsu Province fundamental research key programs. She was nominated for the 2025 Top Ten Advances in IoT Technology, won the National First Prize in the China Collegiate Computing Competition—Network Technology Challenge, and the First Prize in the 2024 Ubiquitous Intelligent Sensing Technology Innovation Application Competition (as advisor).

Report Title: The “Past and Present” of Moiré Pattern Visual Perception

Report introduction: 

       Traditional computer vision perception accuracy is typically limited to the pixel level, making it challenging to achieve “super-resolution” perception of the microscopic 3D world. Achieving ultra-high precision spatial perception remains a constant pursuit. This talk will introduce research on ultra-high precision spatial perception based on “moiré pattern” vision and its extensions. Historically, moiré patterns caused by high-frequency interference in image sensors were considered an obstacle to image quality, and researchers sought to eliminate them. However, Ning’s team took an alternative approach—shifting from “elimination” to “enhancement”—and innovatively explored a theoretical and technical framework for moiré-based visual perception, breaking current resolution limits and ushering in a new era for moiré vision. Applications in intelligent manufacturing, smart healthcare, and other cutting-edge fields will also be shared.

introduction:
       Liu Yimeng, Associate Researcher, obtained her Ph.D. from Northwestern Polytechnical University under Prof. Yu Zhiwen, and is currently with East China Normal University. She works at the intersection of ubiquitous computing, human-machine collaborative computing, and learning sciences. She has years of experience in high-level architecture design and key algorithm research in crowdsensing and human-machine collaboration, as well as rich experience in algorithm design and practical deployment for learning sciences, gamified learning, and intelligent teaching systems. She has published over 20 papers in top conferences/journals such as IEEE TMC, CSCW, IoTJ, ESWA, and KDD, and holds over 20 invention patents and software copyrights. She serves as a reviewer for TMC, TSC, THMS, and other top journals. She leads Shanghai Natural Science Foundation and education research projects and has participated in NSFC and national key R&D programs.

Report Title: Research on Architecture and Key Technologies for Heterogeneous Crowdsensing Systems

Report introduction: 

       Heterogeneous crowdsensing and computing paradigms face challenges such as the lack of a general configurable system architecture and missing core functional modules. This talk introduces a human-machine-thing integrated heterogeneous crowdsensing ubiquitous system architecture (CrowdOS) and its applications in public safety. The architecture addresses task representation and understanding, crowd entity management, real-time interaction, and collaborative scheduling. It offers a comprehensive, modular, and scenario-decoupled framework. Based on CrowdOS and its key technologies, the talk will further present multi-object tracking applications in complex spatiotemporal environments.

introduction:
       Chen Liyue, Ph.D. from Peking University under Prof. Wang Leye, focuses on spatiotemporal data mining and time series analysis. In recent years, he has published around 10 high-level papers in ubiquitous computing-related top venues, including SIGKDD, ICDE, and IEEE TMC (CCF-A). His work has been applied in industrial smart transportation systems. During his Ph.D., he was a core developer of the spatiotemporal forecasting toolbox UCTB, which has received 400+ GitHub stars, contributing to community growth.

Report Title: Urban Spatiotemporal Forecasting via Region Partitioning and Knowledge Transfer

Report introduction: 

       With the rapid development of smart cities, urban sensing networks and multi-source spatiotemporal data systems are improving, enabling better understanding and fine-grained management of urban operations. The performance of urban spatiotemporal forecasting depends not only on the model but also on effective region partitioning and cross-region knowledge transfer. Most existing methods assume manually defined partitions, ignoring their deep impact on tasks. Poor partitioning can severely degrade prediction accuracy. This talk focuses on optimizing region partitioning as the entry point, presenting a full-chain approach from partitioning to knowledge transfer, and discussing applications in different city forecasting scenarios.

introduction:
       Wang Qianru, Ph.D. from Northwestern Polytechnical University under Prof. Guo Bin, was a joint Ph.D. student at Arizona State University, and is now with the School of Computer Science and Technology at Xidian University while also conducting postdoctoral research at City University of Hong Kong. She was selected for the first CCF Ubiquitous Computing Committee Outstanding Doctoral Dissertation Incentive Program, the 2024 Hong Kong Scholars Program, and received NSFC Young Scientist funding. Her research focuses on spatiotemporal prediction for smart cities and continual learning on edge devices. She has published over 10 papers in UbiComp, CSCW, ICDE, TMC, and other CCF-A venues.

Report Title: Spatiotemporal Sequence Prediction for Cities Based on Deep Probabilistic Graph Models

Report introduction: 

       With the rapid development and popularity of IoT and mobile internet, urban spatiotemporal big data with temporal and geographic information have become vital for smart city development. This talk focuses on typical urban applications such as safety risk estimation and user demand forecasting. Addressing challenges such as sparsity, missing data, and dynamic changes in spatiotemporal data, it will introduce deep probabilistic graph-based methods for spatiotemporal prediction and causal impact evaluation, aiming to provide strong theoretical and practical guidance for future research in this field.