ICDM 2020

155 papers
The 2020 IEEE International Conference on Data Mining (ICDM) featured 230 papers addressing diverse topics in machine learning and data science. A significant portion of the research focused on automated heuristic design using large language models, hardware-aware neural architecture search, and the joint optimization of architectures and hyperparameters. Graph-based methodologies were also prominent, including dynamic graph attention for spatio-temporal systems and relational learning on heterogeneous information networks. Interpretability and transparency were central themes, with studies investigating faithfulness in feature attribution and diverse counterfactual explanations for algorithmic recourse. In recommendation systems, research examined the gap between offline metrics and online performance, as well as methods for scalable justification. Application-specific work included geo-spatiotemporal analysis, technical phrase extraction for patent mining, and molecular property prediction using heterogeneous graph neural networks. Prolific authors for this year included Enhong Chen, Qi Liu, and Yanjie Fu, among others.

Top Research Directions

Research directions most represented by ICDM 2020 papers.

Adapting Homogeneous Graph Learning to Heterogeneous Structures 2
Generalizing Influence Maximization Across Diffusion Models 1
Compression-Based Proxies for Kolmogorov Complexity 1
Approximating the Uncomputable Ideals of Kolmogorov Complexity 1
Graph Constrained Fair Division 1
Joint Forecasting and Reconstruction for Sequential Anomaly Detection 1
Adaptive Data Augmentation Driven by Model State 1
Automated Discovery of Optimization Heuristics 1
Aligning Game Mechanics with Player Motivation Profiles 1
Bridging the Topological Semantic Gap in Network Diffusion 1
Causal Discovery via Minimum Description Length 1
Scalable Log Analysis for High Volume Distributed Systems 1
Disentangling Selection and Influence in Social Networks 1
Joint Embedding of Graph Structure and Node Features 1
Scaling Multi-Relational Models Without Parameter Explosion 1

Top Authors

AuthorPapers
Enhong Chen5
Qi Liu5
Di Jin4
Xiangliang Zhang4
Yanjie Fu4
Carlotta Domeniconi3
Christos Faloutsos3
Chuan Shi3
Chuan Zhou3
Guoxian Yu3
Jianhui Ma3
Jiawei Zhang3
Jia Wu3
Jun Wang3
Pengyang Wang3
Philip S. Yu3
Stratis Ioannidis3
Zhenya Huang3
Baojian Zhou2
Bryan Hooi2

Impactful Papers

Papers from ICDM 2020 that shaped research clusters.

Foundational

Influential

Papers

Title
Mining Text Outliers in Document Directories
MixBoost: Synthetic Oversampling using Boosted Mixup for Handling Extreme Imbalance
Modeling Personalized Out-of-Town Distances in Location Recommendation
Multi-Attention 3D Residual Neural Network for Origin-Destination Crowd Flow Prediction
Multi-typed Objects Multi-view Multi-instance Multi-label Learning
Multivariate Time-series Anomaly Detection via Graph Attention Network
NAG: Neural feature aggregation framework for credit card fraud detection
NF-VGA: Incorporating Normalizing Flows into Graph Variational Autoencoder for Embedding Attribute Networks
Online AUC Optimization for Sparse High-Dimensional Datasets
Online Bayesian Sparse Learning with Spike and Slab Priors
Online Optimized Product Quantization
OpenWGL: Open-World Graph Learning
Open-World Class Discovery with Kernel Networks
Opinion Dynamics Incorporating Higher-Order Interactions
Ordinal-Content VAE: Isolating Ordinal-Valued Content Factors in Deep Latent Variable Models
Pairwise Learning for Name Disambiguation in Large-Scale Heterogeneous Academic Networks
Partial Multi-label Learning using Label Compression
Perfect: A Hyperbolic Embedding for Joint User and Community Alignment
Prescriptive Learning for Air-Cargo Revenue Management
Probabilistic Neural-Kernel Tensor Decomposition
Provably Robust Node Classification via Low-Pass Message Passing
Pruning Deep Neural Networks with $\ell_{0}$-constrained Optimization
Quality meets Diversity: A Model-Agnostic Framework for Computerized Adaptive Testing
Random Histogram Forest for Unsupervised Anomaly Detection
Rethinking Dice Loss for Medical Image Segmentation
Rethinking Local Community Detection: Query Nodes Replacement
Robust Collaborative Learning with Noisy Labels
Robust Meta Network Embedding against Adversarial Attacks
Robust Product Rating Rules Against Herding Effects: Theory and Applications
Self-Attentive Classification-Based Anomaly Detection in Unstructured Logs
Semantics-assisted Wasserstein Learning for Topic and Word Embeddings
Semi-Supervised Abductive Learning and Its Application to Theft Judicial Sentencing
Semi-Supervised Partial Multi-Label Learning
Simplifying Reinforced Feature Selection via Restructured Choice Strategy of Single Agent
Soft Graph Matching: Submodular Relaxation and Lovász Extension
Spatiotemporal Hypergraph Convolution Network for Stock Movement Forecasting
STGCN: A Spatial-Temporal Aware Graph Learning Method for POI Recommendation
Stochastic Hard Thresholding Algorithms for AUC Maximization
Structure-based Knowledge Tracing: An Influence Propagation View
Sub-graph Contrast for Scalable Self-Supervised Graph Representation Learning
T2-Net: A Semi-supervised Deep Model for Turbulence Forecasting
TADO: Time-varying Attention with Dual-Optimizer Model
Technical Phrase Extraction for Patent Mining: A Multi-level Approach
The Necessity of Geometrical Representation for Deep Graph Analysis
The Phantom Steering Effect in Q&A Websites
The Relaxed Maximum Entropy Distribution and its Application to Pattern Discovery
TopicOcean: An Ever-Increasing Topic Model With Meta-learning
Toward Model Parallelism for Deep Neural Network based on Gradient-free ADMM Framework
Towards Inspecting and Eliminating Trojan Backdoors in Deep Neural Networks
Toward Unsupervised Graph Neural Network: Interactive Clustering and Embedding via Optimal Transport

« prev Page 3/4 next »