I am a PhD student in the Department of Computer Science & Engineering at University of Notre Dame. I work as a research assistant at Social Sensing Lab under supervision of Prof. Dong Wang. I got my M.S. in Data Science from Indiana University-Bloomington. Before that, I attained my B.S. in Software Engineering from Wuhan University.
Email: yzhang42 [at] nd [dot] edu
My research focuses on the assured artificial intelligence systems with uncertainty quantification in Human-Cyber-Physical systems.
Project 1: Reliable Multimodal Data Fusion from Physical and Human Sensors in Intelligent Transportation Systems (Role: Leader)
- Design a set of novel AI frameworks that establish reliable multi-modal data sensing and fusion to explore the collective power of physical and human sensors.
- Develop a graph neural network framework to accurately forecast the traffic risks in a city by jointly exploring the socialmedia sensing and remote sensing paradigms.
- Propose a novel multi-view learning scheme to accurately identify the risky traffic locations by fusing the semantic and visualfeatures from the heterogeneous social and remote sensing data.
- Keywords: Multi-Modal Data Fusion, Graph Neural Network, Multi-View Learning
- Publication(s): IEEE BigData’18, ACM/IEEE ASONAM’19, IEEE DCOSS’19, ICCPS’20 (Under Review)
Project 2: Migratable Land Usage Classification using Remote Sensing in Smart City (Role: Leader)
- Focus on accurately classifying the land usage of locations in a target city where the ground truth land usage data is not available by leveraging a classification model from a source city where such data is available.
- Propose an adversarial transfer learning framework to translate the satellite images from the target city to the source city for accurate land usage classification.
- The unsupervised nature of proposed framework makes it applicable to the “data drought” problem in similar big data applications where the labeled data from the studied area is unavailable.
- Keywords: Migratable Land Usage Classification, Generative Adversarial Learning, Smart City
- Publication(s): IEEE BigData’19, IPSN’20 (Under Review)
Project 3: Human-AI Hybrid Systems for Disaster Damage Assessment (Role: Member)
- Develop a novel technique using CNN and attention mechanism to automatically assess the severity of damages of disaster scene images collected from remote sensing and social media.
- Propose a human-AI hybrid system that combines active learning and reinforcement learning techniques to acquire human intelligence from crowdsourcing platforms to interpret and further improve the pure CNN-based damage assessment.
- Keywords: Convolutional Neural Network, Reinforcement Learning, Explainable AI
- Publication(s): IEEE ICDCS’19, AAAI’20 (Under Review)
Project 4: Quality-Cost-Aware Task Allocation in Crowd/Social Sensing Applications (Role: Leader)
- Address quality-cost-aware task allocation problem in crowd/social sensing to identify a task allocation strategy (i.e., decide when and where to collect sensing data) to achieve an optimized tradeoff between data quality and sensing cost.
- Build a set of principled online and reinforcement learning frameworks that prove to be effective in reducing sensing cost and improving data quality in real-world social and crowd sensing applications.
- Keywords: Online Learning, Reinforcement Leanring, Task Allocation
- Publication(s): IEEE IPCCC’18, IEEE ICCCN’18, PMC