IEEE BigData 2021
188 papers
The 2021 IEEE International Conference on Big Data (IEEE BigData 2021) featured 235 papers focusing on data efficiency, large-scale simulation, and advanced machine learning architectures. A central theme was the optimization of data annotation processes, with research addressing the human verification bottleneck, selective sampling for label-efficient learning, and the identification of diminishing returns in data labeling. Other studies explored methods to mitigate recency bias in continual learning and techniques for combating noisy labels in deep learning models.
The conference also highlighted applications of big data in societal modeling and public health. This included the development of high-fidelity synthetic populations and the modeling of spatial heterogeneity in infectious disease transmission to inform public health interventions. Technical contributions spanned graph convolutional networks for heterogeneous graphs, privacy-preserving inference, and anomaly detection for high-dimensional sparse data. Prolific contributors to the 2021 edition included researchers such as Da Yan, Philip S. Yu, and Elke A. Rundensteiner.
Top Research Directions
Research directions most represented by IEEE BigData 2021 papers.
Top Authors
Impactful Papers
Papers from IEEE BigData 2021 that shaped research clusters.
Foundational
Training Graph Neural Networks by Graphon Estimation
foundational
Mitigating Oversmoothing in Deep Graph N
Achieving Differential Privacy in Vertically Partitioned Multiparty Learning
foundational
Label Leakage through Gradients in Distr
Influential
Position-based Hash Embeddings For Scaling Graph Neural Networks
influential
Scaling Graph Embeddings Beyond the Memo
Reducing and Exploiting Data Augmentation Noise through Meta Reweighting Contrastive Learning for Te
influential
Hyperparameter Sensitivity in Complex Ma
Forecast-based Multi-aspect Framework for Multivariate Time-series Anomaly Detection
influential
Interpretable Anomaly Detection
Model-Centric and Data-Centric Aspects of Active Learning for Deep Neural Networks
influential
Computational Scalability of Active Lear
Higher-order Structure Based Anomaly Detection on Attributed Networks
influential
Graph Attention Mechanisms for Adaptive
Click-Based Student Performance Prediction: A Clustering Guided Meta-Learning Approach
influential
The Reproducibility Crisis of Private an
Model-Centric and Data-Centric Aspects of Active Learning for Deep Neural Networks
influential
Active Learning for Reducing Data Labeli
CUP: Cluster Pruning for Compressing Deep Neural Networks
influential
Similarity Based Filter Pruning for Mode
Higher-order Structure Based Anomaly Detection on Attributed Networks
influential
Graph Attention Mechanisms for Adaptive
Papers
« prev Page 4/4