IJCNN 2018
232 papers
The 2018 International Joint Conference on Neural Networks (IJCNN) featured 280 papers addressing diverse topics in neural computation and machine learning. Research themes focused heavily on enhancing model robustness and efficiency, with significant work in physics-constrained learning and hierarchical inductive biases for sequence modeling. Affective computing was a prominent area of study, particularly regarding multimodal fusion and bridging domain gaps in speech emotion recognition. Additionally, researchers explored the adaptation of vision foundation models to specialized medical domains and the representation of irregular time series as unordered sets.
Methodological trends included the development of compact visual descriptors for large-scale image retrieval and class-blind pruning techniques for deep neural networks. Practical applications highlighted in the proceedings ranged from power grid contingency anticipation and multiagent path planning to 3D prostate MR segmentation. Prolific authors contributing to this year's edition included Stefan Wermter, Michael Blumenstein, and Badong Chen. Collectively, the research emphasized the integration of domain-specific knowledge into deep learning architectures to improve performance in "in-the-wild" scenarios and specialized technical fields.
Top Research Directions
Research directions most represented by IJCNN 2018 papers.
Top Authors
Impactful Papers
Papers from IJCNN 2018 that shaped research clusters.
Influential
Learning Environmental Sounds with Multi-scale Convolutional Neural Network
influential
Generalization to Unseen Acoustic Enviro
STDP Learning of Image Patches with Convolutional Spiking Neural Networks
influential
Topological Self Organization through Sp
More Realistic and Efficient Face-Based Mobile Authentication using CNNs
influential
Iterative Bounding Box Refinement
Training Deep Neural Networks with Different Datasets In-the-wild: The Emotion Recognition Paradigm
influential
Generalizing Emotion Recognition to Unco
Learning Environmental Sounds with Multi-scale Convolutional Neural Network
influential
Generalization to Unseen Acoustic Enviro
Online Shortest Paths With Confidence Intervals for Routing in a Time Varying Random Network
influential
Online Parameter Tuning for Dynamic Comm
Learning Environmental Sounds with Multi-scale Convolutional Neural Network
influential
Multi-Domain Feature Fusion for Robust S
Spiking Neural Networks Enable Two-Dimensional Neurons and Unsupervised Multi-Timescale Learning
influential
Overcoming the Rate Coding Bottleneck in