IJCNN 2020

440 papers
The 2020 International Joint Conference on Neural Networks (IJCNN) featured 546 papers, with research focusing on domain adaptation and the development of invariant representations. Key themes involved improving model generalization through statistical alignment and unsupervised learning. Knowledge distillation was also a major area of study, particularly regarding collaborative training and dynamic weighting for adaptive model training. Additionally, the conference highlighted advancements in model robustness, such as adversarial purification via generative manifold reconstruction and geometric stability for deep neural networks. Other research directions included continual learning through feature stabilization and efficient optimization for dense prediction. Practical applications presented during the event ranged from traffic signal control using centralized coordinators to multimodal autoencoders for music recommendation and transformer-based architectures for abstractive text summarization. Prolific authors contributing to this year's proceedings included Imran Razzak, Albert Bifet, and Saeeda Naz. The research collectively addressed the challenges of making neural networks more adaptive, resilient, and efficient across diverse data environments.

Top Research Directions

Research directions most represented by IJCNN 2020 papers.

Supervised Dimensionality Reduction 3
Clean Accuracy Degradation in Machine Learning Defenses 3
Mask-Aware Convolutions for Incomplete Data 2
Decoupling Representation and Classifier Learning 2
Dimensionality Reduction via Autoencoder Bottlenecks 2
Mixup Regularization via Linear Sample Interpolation 2
Eliminating Spurious Feature Reliance in Medical Imaging Models 2
Global Context Modeling via Non-Local Spatial Attention 2
Bridging the Discrete and Continuous Reinforcement Learning Divide 2
Stacked Generalization for Heterogeneous Ensemble Learning 2
Safe Reinforcement Learning via Constrained Markov Decision Processes 2
Sample Efficient Reinforcement Learning 2
Incremental Model Updates for Continuous Data Streams 2
Learning in Non-Stationary Data Streams 2
Bypassing the Computational Bottleneck of Adversarial Training 2

Top Authors

AuthorPapers
Imran Razzak8
Albert Bifet5
Andreas Dengel5
Cleber Zanchettin5
Muhammad Shafique5
Alberto Marchisio4
Dongbin Zhao4
Guangquan Zhang4
Jie Lu4
Lina Yao4
Qinru Qiu4
Sarah Monazam Erfani4
Tingwen Liu4
A. K. Qin3
Bin Wang3
Christopher Leckie3
David Macêdo3
Dinh Phung3
Fakhri Karray3
Ivan Y Tyukin3

Impactful Papers

Papers from IJCNN 2020 that shaped research clusters.

Foundational

Influential

Papers

Title
The Variational Deficiency Bottleneck
TimeSAN: A Time-Modulated Self-Attentive Network for Next Point-of-Interest Recommendation
Topological Gradient-based Competitive Learning
Toward Improving the Evaluation of Visual Attention Models: a Crowdsourcing Approach
Towards Best Practice in Explaining Neural Network Decisions with LRP
Towards Improved Deep Contextual Embedding for the identification of Irony and Sarcasm
Towards Online Discovery of Data-Aware Declarative Process Models from Event Streams
Towards Personalized Aesthetic Image Caption
Towards Precise End-to-end Semi-Supervised Human Head Detection Network
Towards Selective Data Enhanced Implicit Discourse Relation Recognition via Reinforcement Learning
Towards Uncovering Feature Extraction From Temporal Signals in Deep CNN: the ECG Case Study
Toward Tag-free Aspect Based Sentiment Analysis: A Multiple Attention Network Approach
Transfer learning in hierarchical dialogue topic classification with neural networks*
TransKP: Transformer based Key-Phrase Extraction
TRec: Sequential Recommender Based On Latent Item Trend Information
TSception: A Deep Learning Framework for Emotion Detection Using EEG
TSCNN: A 3D Convolutional Activity Recognition Network Based on RFID RSSI
TULSN: Siamese Network for Trajectory-user Linking
Two-stage Automatic Image Annotation Based on Latent Semantic Scene Classification
Two-stage Strategy for Small-footprint Wake-up-word Speech Recognition System
Unconstrained Arabic Scene Text Analysis using Concurrent Invariant Points
Underwater object detection using Invert Multi-Class Adaboost with deep learning
Unified Multi-Domain Learning and Data Imputation using Adversarial Autoencoder
Universal Adversarial Perturbations in Epileptic Seizure Detection
Unleashing the Potential of Attention Model for News Headline Generation
Unseen Target Stance Detection with Adversarial Domain Generalization
Unsupervised Clustering through Gaussian Mixture Variational AutoEncoder with Non-Reparameterized Variational Inference and Std Annealing
Unsupervised Deep Imputed Hashing for Partial Cross-modal Retrieval
Unsupervised Features Extracted using Winner-Take-All Mechanism Lead to Robust Image Classification
Unsupervised Learning of Disentangled Location Embeddings
Using Autoencoders for Anomaly Detection in Hard Disk Drives
Variational Clustering: Leveraging Variational Autoencoders for Image Clustering
Visual-to-Semantic Hashing for Zero Shot Learning
Wasserstein Exponential Kernels
Weakly Supervised Object Localization Using Self-Paced Pyramid Adversarial Learning
When Explainability Meets Adversarial Learning: Detecting Adversarial Examples using SHAP Signatures
Who Cried When: Infant Cry Diarization with Dilated Fully-Convolutional Neural Networks
Why Not? Tell us the Reason for Writer Dissimilarity
WW-Nets: Dual Neural Networks for Object Detection
Zero-Shot Object Detection with Textual Descriptions Using Convolutional Neural Networks

« prev Page 9/9