ICDE 2023
290 papers
The 2023 edition of ICDE featured 433 papers, with research themes centered on time series analysis, pattern mining, and the reliability of machine learning models. Key areas of investigation included deep seasonal trend decomposition, spatio-temporal forecasting, and efficient pattern mining in temporal graphs. A notable trend was the focus on uncertainty quantification and calibration for deep networks, alongside methods for resolving parameter interference in heterogeneous federated learning.
Specific contributions included "Sudowoodo," a framework for contrastive self-supervised learning in data integration, and "Saguaro," an edge computing-enabled hierarchical permissioned blockchain. Research also addressed data stream processing, entity graph learning for user targeting, and functional dependency approximation. Prolific authors for this year included Xuemin Lin, Wenjie Zhang, Xiaofang Zhou, and Ye Yuan. The proceedings reflected an ongoing integration of exogenous context into forecasting models and advancements in high-utility pattern mining beyond traditional downward closure constraints.
Top Research Directions
Research directions most represented by ICDE 2023 papers.
Top Authors
Impactful Papers
Papers from ICDE 2023 that shaped research clusters.
Influential
Sequential Recommendation with User Causal Behavior Discovery
influential
Optimizing the Incremental Impact of Rec
Data Imputation for Sparse Radio Maps in Indoor Positioning
influential
Reconstructing Radio Maps from Sparse Fi
UniMatch: A Unified User-Item Matching Framework for the Multi-purpose Merchant Marketing
influential
Dynamic Preference Modeling in Interacti
FedKNOW: Federated Continual Learning with Signature Task Knowledge Integration at Edge
influential
Peer to Peer Consensus in Serverless Fed
MAMDR: A Model Agnostic Learning Framework for Multi-Domain Recommendation
influential
Gradient Alignment for Robust Domain Gen
Jointly Attacking Graph Neural Network and its Explanations
influential
Adversarial Robustness in Graph Neural N
EdgeVisionBench: A Benchmark of Evolving Input Domains for Vision Applications at Edge
influential
Domain Adaptation with Category Shift