Research Areas

Big Data Analytics and Management (BDAM)

The BDAM research group aims to facilitate secure, effective, and efficient use and management of big data under a wide variety of hardware, software, and organizational settings. The research topics include data analytics, blockchain, databases, data privacy and security, query processing, and graph/social/spatial data management.

Faculty Involved:

• Prof. BAI, Jiaxin
• Prof. CHOI, Byron Koon Kau
• Prof. HUANG, Xin
• Prof. MA, Jing
• Dr. PIAO, Chengzhi
• Prof. XU, Jianliang
• Prof. YANG, Renchi

Funded Research and Consultancy Projects in the Past Few Years:

Querying Large Networks of Cryptocurrencies
	Staff	Prof. CHOI, Byron Koon Kau
	Objectives	Propose and analyze novel flow queries on large networks in static and streaming settings Investigate scalable and efficient querying algorithms for flow queries Discover large transactions of cryptocurrencies (Blockchains) and sudden bursts of traffic in road networks
	Grant	General Research Fund (GRF)

Federated Graph Management and Querying: Subgraphs, Keywords, and Privacy
	Staff	Prof. HUANG, Xin
	Objectives	Propose efficient algorithms for federated graph management and query processing to support fundamentally useful subgraph search and counting queries, including the k-core search and triangle counting; Design federated attributed graph analytics algorithms for spatial-temporal community search and keyword search; Develop new techniques for enhancing privacy protection for federated subgraph search and counting, and also a paradigm solution for differentially private federated graph analytics; More Collaborate with industrial partners to develop and implement a prototype system of federated graph database based on all integrated techniques.
	Grant	Collaborative Research Fund (CRF)

Conversation-based Reasoning for Explainable Fake News Detection via Weakly Supervised Evidence Detection
	Staff	Prof. MA, Jing
	Objectives	Develop a robust explainable fake news detection system, which not only show the veracity of breaking news but also give human-understandable explanation to support the judgement. Design and develop weakly supervised framework that are capable of digesting massive web data and inferring evidences with minimal annotations. Build new benchmarks for explainable fake news detection containing low-resource domains, and propose both qualitative and quantitative criteria for evaluation.
	Grant	Early Career Scheme (ECS)

Vector Similarity Search in High-Dimensional Spaces
	Staff	Prof. XU, Jianliang
	Objectives	Develop novel proximity graph (PG)-based indexing methods to reduce the time complexity of single-vector similarity search in high-dimensional spaces. Propose a margin-allowed PG framework for efficient and accurate multi-vector similarity search in high-dimensional spaces. Design a new hierarchical PG framework to minimize I/O costs and enable efficient vector similarity search in large-scale databases. More Conduct comprehensive theoretical analysis and empirical studies to evaluate the effectiveness of the proposed techniques and algorithms.
	Grant	General Research Fund (GRF)

Realtime Relevance Search over Massive Attributed Graphs
	Staff	Prof. YANG, Renchi
	Objectives	Design a novel relevance measure specifically for attributed graphs that learns personalized weights for combining topology and attribute information. Propose optimization techniques to allow for efficient parameter learning on massive attributed graphs. Create new index structures based on network embedding candidate sets to enable real-time relevance search on large static graphs.More Formulate an online bidirectional search algorithm to handle real-time top-k query processing. Explore dynamic network embedding and incremental index-update methods to extend real-time search capabilities to large dynamic graphs.
	Grant	Early Career Scheme (ECS)