Neural Information Processing: 28th International Conference, ICONIP 2021, Sanur, Bali, Indonesia, December 8–12, 2021, Proceedings, Part III (Lecture Notes in Computer Science Book 13110)

The four-volume proceedings LNCS 13108, 13109, 13110, and 13111 constitutes the proceedings of the 28th International Conference on Neural Information Processing, ICONIP 2021, which was held during December 8-12, 2021. The conference was planned to take place in Bali, Indonesia but changed to an online format due to the COVID-19 pandemic. The total of 226 full papers presented in these proceedings was carefully reviewed and selected from 1093 submissions. The papers were organized in topical sections as follows: Part I: Theory and algorithms; Part II: Theory and algorithms; human centred computing; AI and cybersecurity; Part III: Cognitive neurosciences; reliable, robust, and secure machine learning algorithms; theory and applications of natural computing paradigms; advances in deep and shallow machine learning algorithms for biomedical data and imaging; applications; Part IV: Applications. Preface Organization Contents – Part III Cognitive Neurosciences A Novel Binary BCI Systems Based on Non-oddball Auditory and Visual Paradigms 1 Introduction 2 Materials and Methods 2.1 Participants and Data Acquisition 2.2 Experimental Paradigms and Task Definitions 2.3 Experiment I: Non-oddball Visual Cue (NV) 2.4 Experiment II: Non-oddball Auditory Cue (NA) 3 Data Analysis and Performance Evaluations 3.1 Linear Classifier Model 3.2 ERP Data Augmentation 3.3 Classification with CNN 4 Results 4.1 ERP Responses in Non-oddball Visual/Auditory Paradigms 4.2 Decoding Accuracy of Non-oddball Visual/Auditory Paradigm 5 Discussion 6 Conclusion References A Just-In-Time Compilation Approach for Neural Dynamics Simulation 1 Introduction 2 Results 2.1 The Overview of the Framework 2.2 Stage 1: Defining Numerical Solvers 2.3 Stage 2: Building Brain Objects 2.4 Stage 3: Simulation with JIT Acceleration 2.5 Stage 4: Analysis and Visualization of Simulation Output 3 Comparison of Running Efficiency 4 Conclusion References STCN-GR: Spatial-Temporal Convolutional Networks for Surface-Electromyography-Based Gesture Recognition 1 Introduction 2 Related Work 3 Spatial-Temporal Convolutional Networks 4 Experiments 4.1 Datasets and Settings 4.2 Comparison Results 4.3 Ablation Study 4.4 Visualization of the Learned Graphs 5 Conclusion References Gradient Descent Learning Algorithm Based on Spike Selection Mechanism for Multilayer Spiking Neural Networks 1 Introduction 2 SNN Architecture and spiking Neuron Model 2.1 SNN Architecture 2.2 Spiking Neuron Model 3 Spike Selection Mechanism and learning Rules of the improved Algorithm 3.1 Spike Selection Mechanism 3.2 Learning Rules of the improved Algorithm 4 Simulation Results 4.1 Parameter Settings 4.2 Spike Train Learning 5 Conclusion References Learning to Coordinate via Multiple Graph Neural Networks 1 Introduction 2 Background 2.1 Dec-POMDP 2.2 Value-Decomposition Multi-agent RL 2.3 Graph Convolutional Networks 3 MGAN 3.1 Embedding Generation via Graph Networks 3.2 MGAN Mixing Network 3.3 Loss Function 4 Experiments 4.1 Settings 4.2 Validation 4.3 Graph Embedding and Weight Analysis 5 Conclusion References A Reinforcement Learning Approach for Abductive Natural Language Generation 1 Introduction 2 Related Work 2.1 Abductive Reasoning Generation Tasks 2.2 Reinforcement Learning in Natural Language Generation 3 Methodology 3.1 Problem Formulation 3.2 Initial LM Training 3.3 Reward Model Training 3.4 Desired LM Training 4 Experiment and Discussion 4.1 Automatic Evaluation 4.2 Learning Curves 4.3 Effect of the Teacher Model 4.4 Human Evaluation 4.5 Showcase Analysis 5 Conclusion References DFFCN: Dual Flow Fusion Convolutional Network for Micro Expression Recognition 1 Introduction 2 Related Works 3 Proposed Method 3.1 Pre-processing 3.2 GAN-Based Synthetic Data Augmentation 3.3 Model Architecture 4 Experiments 4.1 Datasets and Evaluation Metrics 4.2 Implementation Details 4.3 Results and Discussion 5 Conclusion References AUPro: Multi-label Facial Action Unit Proposal Generation for Sequence-Level Analysis 1 Introduction 2 Related Work 3 Method 3.1 Task Setting 3.2 AUPro Architecture 3.3 Training of AUPro 4 Evaluation 4.1 Experimental Settings 4.2 AUPro vs. TAG-Based Method 4.3 Qualitative Results 5 Conclusion References Deep Kernelized Network for Fine-Grained Recognition 1 Introduction 2 Related Work 3 Study Design 3.1 Kervolution 3.2 Kernelized Dense Layer 3.3 Support Vector Machine 4 Experiments 4.1 Datasets 4.2 Models Architecture and Training Process 4.3 Ablation Study 5 Conclusion References Semantic Perception Swarm Policy with Deep Reinforcement Learning 1 Introduction 2 Background 2.1 Partially Observable Markov Games 2.2 -VAE 2.3 Distributed Swarm Systems 3 Approach 3.1 Overall Design of SPSP 3.2 Semantic Perception 3.3 Training Method of SPSP 4 Results 4.1 Task Description 4.2 Simulation Results 4.3 Real-World Experiment Results 5 Conclusion References Reliable, Robust, and Secure Machine Learning Algorithms Open-Set Recognition with Dual Probability Learning 1 Introduction 2 Related Work 3 Proposed Method 3.1 Dual Probability Formulation 3.2 Architecture 3.3 Training 3.4 Testing 4 Experiments and Results 4.1 Experimental Setup 4.2 Ablation Analysis 4.3 Comparison with State-of-Art Results 5 Conclusion References How Much Do Synthetic Datasets Matter in Handwritten Text Recognition? 1 Introduction 2 Related Work 3 Our Method – Synthetic Dataset Generation 3.1 Synthetic Schema-Based Character Generator 3.2 Autoencoder-Based Generators 4 Research Design 4.1 Experimental Database Preprocessing 4.2 Training with Synthetic and Handwritten Data 4.3 Base Classifier for Character Recognition 5 Experiments and Results 6 Conclusions References PCMO: Partial Classification from CNN-Based Model Outputs 1 Introduction 2 Related Work 3 Background 3.1 The Pattern of the CNN-Based Model Outputs 3.2 Dempster-Shafer Theory 4 Proposed Method 5 Experiments 5.1 Experiment Protocol 5.2 Criteria 5.3 Evaluation of the PCMO Method 5.4 Practical Usage of the PCMO Method 6 Conclusion References Multi-branch Fusion Fully Convolutional Network for Person Re-Identification 1 Introduction 2 Related Works 3 Proposed Method 3.1 Network Architecture 3.2 Training and Loss Functions 4 Experiments 4.1 Datasets and Evaluation Protocols 4.2 Implementation Details 4.3 Comparison with State-of-the-Art Methods 4.4 Further Analysis and Discussions 5 Conclusion References Fast Organization of Objects' Spatial Positions in Manipulator Space from Single RGB-D Camera 1 Introduction 2 Related Work 2.1 Object Extraction from Cluster Environment 2.2 Objects Contours Reconstructed (Eye-to-Hand Calibration) 2.3 Reconstruction Error Elimination 3 The Proposed Method 3.1 DP-DROPA (Density Peaks Based Distance-Restricted Outlier Point Adjustment) 3.2 Object Extraction from Cluster Environment and Object Contours Reconstruction 3.3 DP-DROPA Based 3D Contour Outliers Elimination 4 Experiments 4.1 Experiment Setting 4.2 Object Extraction and Contour Points Selection 4.3 Object Reconstruction 4.4 Time Cost 4.5 Organization of the Objects' Spatial Positions from a Single RGB-D Image 4.6 Discussion 5 Conclusion References EvoBA: An Evolution Strategy as a Strong Baseline for Black-Box Adversarial Attacks 1 Introduction 2 Related Work 3 The Method 3.1 Notation and Threat Model 3.2 Proposed Algorithm 4 Experiments 4.1 Experimental Setup 4.2 Results on MNIST 4.3 Results on CIFAR-10 4.4 Results on ImageNet 5 Conclusion References A Novel Oversampling Technique for Imbalanced Learning Based on SMOTE and Genetic Algorithm 1 Introduction 2 Related Work 2.1 Oversampling Methods 2.2 Genetic Algorithm 3 The Proposed GA-SMOTE Method 4 Experiments 4.1 Datasets Analysis 4.2 Evaluation Metrics 4.3 Compare Methods and Parameter Settings 4.4 Algorithm Comparison and Result Analysis 5 Conclusion References Dy-Drl2Op: Learning Heuristics for TSP on the Dynamic Graph via Deep Reinforcement Learning 1 Introduction 2 Related Work 3 The Proposed Neural Network Model 3.1 Input Preprocessing 3.2 Policy Network Parametrization 3.3 Computational Experiment 3.4 Training Procedure 4 Performance Evaluation 4.1 Hyperparameters 4.2 Comparison of Optimal Performance 4.3 Visualization 5 Conclusion References Multi-label Classification of Hyperspectral Images Based on Label-Specific Feature Fusion 1 Introduction 2 The Proposed Algorithm 2.1 Construction of Distance Mapping Feature Between Instances 2.2 Construction of Linear Representation Mapping Features Between Instances 2.3 Construction of Band Clustering Mapping Feature 2.4 Construction of Classification Model 3 Experimental Results and Analysis 3.1 The Experimental Setup 3.2 Multi-label Algorithm Comparison 4 Conclusion References A Novel Multi-scale Key-Point Detector Using Residual Dense Block and Coordinate Attention 1 Introduction 2 Proposed Method 2.1 Network Architectures 2.2 The Feature Fusion Module with Proposed RDBCA 2.3 Scale Assignment 2.4 Label Assignment 2.5 Loss Function 2.6 Inference 3 Experiments and Results 3.1 Training Details 3.2 Without NKS vs. with NKS 3.3 Proposed RDB vs. Original FPN 3.4 Comparison Results on VOC2007 4 Conclusion References Alleviating Catastrophic Interference in Online Learning via Varying Scale of Backward Queried Data 1 Introduction 2 Related Works 3 Conceptualization 3.1 Stochastic Gradient Descent 3.2 Backward Query 3.3 Proposed Concept 4 Experiment 4.1 Experimental Settings 4.2 Experiment with SGD and the Proposed Method 5 Conclusion References Construction and Reasoning for Interval-Valued EBRB Systems 1 Introduction 2 Problem Formulation 3 Interval Extended Belief Rule-Based System 3.1 Construction of IEBRB 3.2 Reasoning of IEBRB 3.3 New Individual Matching Degree 3.4 Determine Attribute Weights 3.5 Framework 4 Experiment and Analysis 4.1 Interval-Valued Data Problem Cases 4.2 Real-Valued Data Problem Cases 5 Conclusion References Theory and Applications of Natural Computing Paradigms Brain-mimetic Kernel: A Kernel Constructed from Human fMRI Signals Enabling a Brain-mimetic Visual Recognition Algorithm 1 Introduction 1.1 Related Work 1.2 Contributions 2 Materials and Methods 2.1 Participant 2.2 MRI Data Acquisition 2.3 MRI Preprocessing 2.4 Estimating Activation Coefficients of image-Induced fMRI Signals 2.5 Image Preprocessing 2.6 KCCA-SVM 3 Results 4 Discussion References Predominant Sense Acquisition with a Neural Random Walk Model 1 Introduction 2 Related Work 3 Framework of the System 3.1 Pre-processing 3.2 Producing Embedding 3.3 Building a Graph with an Adjacency Matrix 3.4 Predicting Categories and Propagation 4 Applying DSS to Text Categorization 5 Experiments 6 Conclusion References Processing-Response Dependence on the On-Chip Readout Positions in Spin-Wave Reservoir Computing 1 Introduction 2 Spin-Wave Reservoir Chip 3 Numerical Analysis of Reservoir Response 3.1 Learning in Sinusoidal and Square Wave Distinction and Its Processing Results 3.2 Origins of the Latency and the Time Lag 4 Conclusion References Advances in Deep and Shallow Machine Learning Algorithms for Biomedical Data and Imaging A Multi-task Learning Scheme for Motor Imagery Signal Classification 1 Introduction 2 Related Works 3 Proposed Methodology 3.1 Preprocessing 3.2 Learning Architectures 4 Dataset, Results and Discussion 4.1 Dataset 4.2 Comparison of Results of Single Task and Multi-task Models 4.3 Comparison of Different Multi-task Models 4.4 Comparison of Multi-task Model with State-of-the-Art Methods 4.5 Effect of Different Hyperparameters 5 Conclusion References An End-to-End Hemisphere Discrepancy Network for Subject-Independent Motor Imagery Classification 1 Introduction 2 Methods 2.1 Data Description and Preprocessing 2.2 Proposed Hemisphere Discrepancy Network 3 Experiments and Results 3.1 Implementation Details 3.2 Baseline Models 3.3 Results 3.4 Ablation Study 3.5 Limitations and Future Work 4 Conclusion References Multi-domain Abdomen Image Alignment Based on Joint Network of Registration and Synthesis 1 Introduction 2 Related Work 3 Proposed Method 3.1 The Overview of RSNet 3.2 The Optimization of the Synthesizer in the Joint Training 3.3 The Optimization of the Register in the Joint Training 4 Experiments 4.1 Datasets and Implementation Details 4.2 Comparison Analysis 4.3 Ablation Studies 5 Conclusions References Coordinate Attention Residual Deformable U-Net for Vessel Segmentation 1 Introduction 2 Method 2.1 CARDU-Net 2.2 Residual Blocks with Dropblock 2.3 Deformable Convolution 3 Experiments and Results 4 Conclusion References Gated Channel Attention Network for Cataract Classification on AS-OCT Image 1 Introduction 2 Related Work 2.1 Cataract Classification 2.2 Attention Mechanism 3 Method 3.1 Revisiting of Channel Attention 3.2 Gated Channel Attention Block 4 Experiments 4.1 Dataset and Evaluation Measures 4.2 Comparison with State-of-art Attention Attention Blocks 4.3 Ablation Study 5 Conclusion References Overcoming Data Scarcity for Coronary Vessel Segmentation Through Self-supervised Pre-training 1 Introduction 2 Related Work 3 Methods 3.1 The Jerman Vesselness Filter 3.2 Volume Masking 3.3 Segmentation Neural Network 3.4 Self-supervised Training and Validation 4 Results and Discussion 5 Conclusions and Future Work References Self-Attention Long-Term Dependency Modelling in Electroencephalography Sleep Stage Prediction 1 Introduction 2 Motivation 3 Method 3.1 Embedder 3.2 Sequence Encoder Stack 4 Experiments 4.1 Datasets 4.2 Preparation 4.3 Model Setup 4.4 Training 4.5 Protocol 5 Results and Discussion 6 Conclusion References ReCal-Net: Joint Region-Channel-Wise Calibrated Network for Semantic Segmentation in Cataract Surgery Videos 1 Introduction 2 Related Work 3 Methodology 4 Experimental Settings 5 Experimental Results 6 Conclusion References Enhancing Dermoscopic Features Classification in Images Using Invariant Dataset Augmentation and Convolutional Neural Networks 1 Introduction 2 Importance of the Invariant Dataset Augmentation 3 Dermoscopic Screening Methodologies and Image Datasets 4 Available Pretrained CNN Networks and Their Parameters 5 General Method Description 6 Results 7 Conclusions References Ensembles of Randomized Neural Networks for Pattern-Based Time Series Forecasting 1 Introduction 2 Forecasting Model 3 Ensembling 4 Experiments and Results 5 Conclusion References Grouped Echo State Network with Late Fusion for Speech Emotion Recognition 1 Introduction 2 Proposed SER Model 2.1 Feature Extraction 2.2 Grouped ESN 2.3 Dimension Reduction with PCA 2.4 Reservoir Model Space and Readout 2.5 Bayesian Hyperparameter Optimization 3 Experimental Setup and Results 3.1 SAVEE 3.2 FAU Aibo Emotion Corpus 4 Discussion 5 Conclusion References Applications MPANet: Multi-level Progressive Aggregation Network for Crowd Counting 1 Introduction 2 Related Works 2.1 Single-Column Based Approaches 2.2 Multi-column Based Approaches 3 Approach 3.1 Overview 3.2 Aggregation Refinement Module 3.3 Multi-scale Aware Module 3.4 Semantic Attention Module 3.5 Loss Function 4 Experiments 4.1 Implementation Details 4.2 Evaluation Metric 4.3 Comparison with the State-of-the-art 4.4 Ablation Study 5 Conclusion References AFLLC: A Novel Active Contour Model Based on Adaptive Fractional Order Differentiation and Local-Linearly Constrained Bias Field 1 Introduction 2 Methodology 2.1 Adaptive Fractional Order Differentiation 2.2 Local-Linearly Constrained Bias Field 2.3 Model Optimization 3 Experiment and Analysis 3.1 Experimental Analysis 3.2 Robustness Analysis for Initial Condition 3.3 The Validity Analysis of the Definition of Adaptive Order 4 Conclusion References DA-GCN: A Dependency-Aware Graph Convolutional Network for Emotion Recognition in Conversations 1 Introduction 2 Related Work 3 Task Definition 4 Proposed Approach 4.1 Utterance Reader 4.2 Utterance Encoder 4.3 Dependency-Aware Graph Layer 4.4 Decoder 5 Experimental Settings and Result Discussions 5.1 Datasets 5.2 Experimental Details 5.3 Baselines 5.4 Overall Results 5.5 Effect of DA-graph Layer Number 5.6 Ablation Study 6 Conclusion References Semi-supervised Learning with Conditional GANs for Blind Generated Image Quality Assessment 1 Introduction 2 Related Work 3 Our Proposal 3.1 The Painter GAN 3.2 Quality Evaluation Model 3.3 Evaluation and Optimization with Score 4 Experimental Results 4.1 Datasets 4.2 Implementation Details 4.3 Supplementary of Evaluation System 4.4 Optimization of Generated Results 4.5 Discussion 5 Conclusion References Uncertainty-Aware Domain Adaptation for Action Recognition 1 Introduction 2 Related Work 3 Methodology 3.1 Domain-Invariant Feature Extractor 3.2 Bayesian Discriminator 3.3 Uncertainty-Aware Attention Mechanism 3.4 Bayesian Classifier 3.5 Loss Function 4 Experiments 4.1 Dataset 4.2 Implementation Details 4.3 Comparison with State-of-the-Art Methods 4.4 Ablation Study 4.5 Visulization 5 Conclusion References Free-Form Image Inpainting with Separable Gate Encoder-Decoder Network 1 Introduction 2 Related Work 2.1 Image Inpainting 2.2 Convolution Method 3 Approach 3.1 Network Structure 4 Experiment 4.1 Experiment Setup 4.2 Quantitative Comparison 4.3 Qualitative Comparison 5 Conclusion References BERTDAN: Question-Answer Dual Attention Fusion Networks with Pre-trained Models for Answer Selection 1 Introduction 2 Related Work 3 Proposed Model 3.1 Task Description 3.2 Overview of Proposed Model 3.3 BERT Encoder 3.4 Cross Attention Between Question and Answer 3.5 Dual Attention Fusion Mechanism 4 Experimental Setup 4.1 DataSet 4.2 Training and Hyperparameters 4.3 Results and Analysis 4.4 Ablation Study 5 Conclusion References Rethinking the Effectiveness of Selective Attention in Neural Networks 1 Introduction 2 Related Work 3 Methodology 3.1 Preliminary 3.2 Rethinking the Effectiveness of Selective Attention in SKNet 3.3 Elastic-SKNet 3.4 Architecture Search for Elastic-SKNet 4 Experiments 4.1 Experimental Settings 4.2 ImageNet Classification Results 4.3 Performance on Auto-ESKNets 4.4 MS-COCO Object Detection Results 5 Conclusion References An Attention Method to Introduce Prior Knowledge in Dialogue State Tracking 1 Introduction 2 Related Work 3 Introducing Named Entity Information 3.1 Extracting Named Entity Information 3.2 Formulating Named Entity Information 4 Approach 4.1 Problem Statement 4.2 Encoding Dialogue Turn and Named Entity Information 4.3 Slot-Dialogue and Slot-Entity Attention 4.4 Historical Context Encoder and Slot-Context Attention 4.5 Outputs and Training Criteria 5 Experiment 5.1 Dataset 5.2 Metrics 5.3 Baselines 5.4 Experiment Settings 5.5 Experiment Results 5.6 Attention Visualization 5.7 Effectiveness of the Employed Named-Entity Recognition Toolkit 5.8 Effectiveness of Dynamic Gating Between Prior Entity Information and Dialogue 5.9 Effectiveness of Prior Entity Information 6 Conclusion References Effect of Input Noise Dimension in GANs 1 Introduction 2 Preliminaries 2.1 DCGAN 2.2 WGAN-GP 2.3 Measures to Evaluate GANs 3 Experiments with Input Noise for GANs 3.1 Results for Synthetic Data-Set 3.2 Results for MNIST 3.3 Results for CelebA 32 3.4 Results for CelebA 64 4 Insights and Future Work 5 Conclusion References Wiper Arm Recognition Using YOLOv4 1 Introduction 2 Prior Work 2.1 Object Recognition 2.2 Object Detection 3 Proposed Method 3.1 Dataset Construction 3.2 Model Training - Data Augmentation and DropBlock Regularization 3.3 YOLOv4 Object Detection 4 Results and Discussion 4.1 Experimental Setup 4.2 Impact of the Proposed Method Under Several Conditions 4.3 Comparison of Different Algorithms with Different Input Resolution at 8000thIteration 5 Conclusions and Recommendations References Context Aware Joint Modeling of Domain Classification, Intent Detection and Slot Filling with Zero-Shot Intent Detection Approach 1 Introduction 2 Related Work 3 Proposed Methodology 3.1 Semantic Module 3.2 Context Module 3.3 Domain Module and Intent Module 3.4 Slot Module 3.5 Zero-Shot Intent Module 4 Implementation Details 4.1 Dataset 4.2 Experimental Setup 5 Results and Analysis 6 Conclusion and Future Work References Constrained Generative Model for EEG Signals Generation 1 Introduction 2 Methodology 2.1 WaveGAN 2.2 SR and AR 3 Experiment and Results 3.1 Selection of Weight Parameter 3.2 Ablation Study 3.3 Comparison with State-of-the-Art Model 4 Conclusion References Top-Rank Learning Robust to Outliers 1 Introduction 2 Related Work 2.1 Top-Rank Learning with Deep Learning 2.2 Outlier Estimation 3 Robust Top-Rank Deep Neural Network 3.1 Top-Rank Learning 3.2 Top-Rank Learning Robust to Outlier 3.3 Combination with Representation Learning 4 Experiment on Artificial Datasets 5 Experiment on Medical Image Diagnosis 5.1 Dataset 5.2 Model Setup 5.3 Comparative Methods 5.4 Quantitative Evaluation by Pos@top 5.5 Quantitative Evaluation of Ranking Stability 6 Conclusion References Novel GAN Inversion Model with Latent Space Constraints for Face Reconstruction 1 Introduction 2 Related Work 2.1 Establishment of GAN Inversion 2.2 Face Recognition Network 2.3 Interpretability and Decoupling Effect of Latent Space 3 Approach 4 Experiments 5 Conclusion References Edge Guided Attention Based Densely Connected Network for Single Image Super-Resolution 1 Introduction 2 Related Work 3 Proposed Method 3.1 Neural Attention 3.2 Edge Generation 3.3 Network Structure 3.4 Loss Function 4 Experiments 4.1 Datasets and Evaluation Metrics 4.2 Implementation Details 4.3 Comparisons with State-of-The-Arts 4.4 Ablation Studies 5 Conclusion References An Agent-Based Market Simulator for Back-Testing Deep Reinforcement Learning Based Trade Execution Strategies 1 Introduction 2 DRL for Optimal Trade Execution 2.1 Preliminaries 2.2 Problem Formulation 2.3 Architecture 3 A Real-Time Simulation Back-Testing Platform 3.1 ABIDES 4 Methodology 4.1 Data Sources 4.2 Experimental Methodology and Settings 4.3 Algorithms 4.4 Configuration for the Simulation Back-Testing Platform 4.5 Main Evaluation and Back-Testing 5 Conclusion and Future Work References Looking Beyond the Haze: A Pyramid Fusion Approach 1 Introduction 2 Related Work and Background 3 Haze Imaging Model 4 Proposed Algorithm 4.1 Feature Extraction 4.2 Image Fusion 5 Experimental Results 5.1 Parameter Settings and Simulation Environment 5.2 Performance Analysis 6 Conclusion and Future Work References DGCN-rs: A Dilated Graph Convolutional Networks Jointly Modelling Relation and Semantic for Multi-event Forecasting 1 Introduction 2 Related Work 3 Methodology 3.1 Preliminaries 3.2 GCN for Relation and Semantic Embedding 3.3 Dilated Casual Convolutions for Temporal Embedding 3.4 Multi-event Forecasting 4 Experiments and Results 4.1 Datasets and Settings 4.2 Baselines 4.3 Experiments Results 4.4 Sensitivity Analysis 5 Conclusion References Training Graph Convolutional Neural Network Against Label Noise 1 Introduction 2 Related Work 2.1 Graph Convolutional Neural Network 2.2 Learning with Noisy Labels 3 Preliminary 4 Methodology 4.1 Graph Construction with Super-Nodes 4.2 Label Correction and Dynamic Graph Adjustment 4.3 Training Iterations 5 Experiments 5.1 Datasets 5.2 Baselines 5.3 Experimental Setup 5.4 Results 5.5 Ablation Studies 6 Conclusions References An LSTM-Based Plagiarism Detection via Attention Mechanism and a Population-Based Approach for Pre-training Parameters with Imbalanced Classes 1 Introduction 2 Long Short-Term Memory (LSTM) 3 LSTM-AM-ABC Approach 3.1 Pre-processing 3.2 Word Embedding 3.3 Model Construction 4 Experiment and Analysis 4.1 Corpus 4.2 Metrics 4.3 Result 5 Conclusions References Author Index The three volume set LNCS 8226, LNCS 8227, and LNCS 8228 constitutes the proceedings of the 20th International Conference on Neural Information Processing, ICONIP 2013, held in Daegu, Korea, in November 2013. The 180 full and 75 poster papers presented together with 4 extended abstracts were carefully reviewed and selected from numerous submissions. These papers cover all major topics of theoretical research, empirical study and applications of neural information processing research. The specific topics covered are as follows: cognitive science and artificial intelligence; learning theory, algorithms, and architectures; computational neuroscience and brain imaging; vision, speech and signal processing; control, robotics and hardware technologies; and novel approaches and applications.