Neural Information Processing

Name: Neural Information Processing | 27th International Conference, ICONIP 2020, Bangkok, Thailand, November 23-27, 2020, Proceedings, Part III
Brand: Springer
Price: 96.29 EUR
Availability: OnlineOnly

27th International Conference, ICONIP 2020, Bangkok, Thailand, November 23-27, 2020, Proceedings, Part III

Haiqin Yang Kitsuchart Pasupa Andrew Chi-Sing Leung James T. Kwok Jonathan H. Chan Irwin King(Editor)

Springer (Publisher)

Published on 18. November 2020

XXV, 643 pages

E-Book

PDF with digital watermarking

System requirements

978-3-030-63836-8 (ISBN)

€96.29incl. 7% vat

System requirements

for PDF with digital watermarking

E-Book Single Licence

Available for download

Description

More details

Other editions

Content

Intro
Preface
Organization
Contents - Part III
Biomedical Information
Classification of Neuroblastoma Histopathological Images Using Machine Learning
1 Introduction
2 Methodology
2.1 Dataset
2.2 Experimental Setup
3 Results
4 Discussion
5 Conclusion
References
Data Mining ENCODE Data Predicts a Significant Role of SINA3 in Human Liver Cancer
1 Introduction
2 Methods
3 Results
3.1 Co-location of TFBS Analysis
3.2 HBV-transfection Effect on Histone Modification
4 Discussion
References
Diabetic Retinopathy Detection Using Multi-layer Neural Networks and Split Attention with Focal Loss
1 Introduction
2 Related Work
3 Methodology
3.1 Proposed Approach
3.2 Dataset
3.3 Data Pre-processing
3.4 Metrics for Performance Evaluation
4 Experimental Results and Discussion
5 Analysis
6 Conclusion
References
Enhancer-DSNet: A Supervisedly Prepared Enriched Sequence Representation for the Identification of Enhancers and Their Strength
1 Introduction
2 Materials and Methods
3 Proposed Enhancer-DSNet Methodology
4 Benchmark Dataset
5 Evaluation Metrics
6 Experimental Setup and Results
6.1 Results
7 Conclusion
References
Machine Learned Pulse Transit Time (MLPTT) Measurements from Photoplethysmography
1 Introduction
2 Design
2.1 Machine Learned Pulse Transit Time
2.2 HeartPy Pulse Transit Time
3 Analysis
3.1 Methods
4 Results
4.1 BIDMC Patient 6 Drilldown
5 Discussion
5.1 Limitations
6 Conclusion
References
Weight Aware Feature Enriched Biomedical Lexical Answer Type Prediction
1 Introduction
2 Related Work
3 Methodology
3.1 Pre-processing
3.2 Feature Selection
3.3 LAT Prediction
4 Experimental Study
4.1 Dataset and Experiment Configuration
4.2 Evaluation Metrics and Baseline
4.3 Results and Discussion
5 Conclusion and Future Work
References
Neural Data Analysis
Decoding Olfactory Cognition: EEG Functional Modularity Analysis Reveals Differences in Perception of Positively-Valenced Stimuli
1 Introduction
2 Methodology
2.1 Recruitment of Participants
2.2 Experimental Protocol
2.3 EEG Data Acquisition and Preprocessing
2.4 EEG Data Analysis - Community Detection
3 Results
3.1 Partitioning the Functional Connectivity Networks
3.2 Inter-module Metrics
3.3 Intra-module Metrics
4 Discussion
4.1 Alterations in the Inter-module Metrics in the Community Functional Connectivity Organisation
4.2 Variations in the Intra-module Metrics in the Community Functional Connectivity Organisation
5 Conclusions
A Appendix: Anatomical Segregation of ROIs in Communities
References
Identifying Motor Imagery-Related Electroencephalogram Features During Motor Execution
1 Introduction
2 Method
2.1 Experiment and Preprocessing Procedure
2.2 Data Analysis
3 Result
3.1 Checking Muscle Activity During Motor Imagery EMG
3.2 Difference Between Motor Imagery and Execution EEG
3.3 Classification Accuracy a Single Trial
4 Conclusion
References
Inter and Intra Individual Variations of Cortical Functional Boundaries Depending on Brain States
1 Introduction
2 Method
2.1 Material and Preprocessing
2.2 Feature of Boundary and the Index Measuring Boundary Variation
2.3 Statistical Differences of Boundary Variation Distribution
2.4 Inter-subject Boundary Variation
2.5 Intra-subject Boundary Variation
2.6 Comparison of Boundary Variation Between Inter- and Intra-subject
3 Result
3.1 Tendency of Inter- and Intra-subject Boundary Variation
3.2 Rankings of Inter- and Intra-subject Boundary Variation on Statistics
3.3 Comparison of Boundary Variation Between Inter- and Intra-subject
4 Conclusion
References
Phase Synchronization Indices for Classification of Action Intention Understanding Based on EEG Signals
1 Introduction
2 Materials and Method
2.1 Subjects
2.2 Experimental Paradigm
2.3 Data Collection and Preprocessing
2.4 Phase Synchronization Indices
3 Results
3.1 Action Intention Understanding Classification
3.2 Brain Network Statistics
4 Discussion
5 Conclusion
References
The Evaluation of Brain Age Prediction by Different Functional Brain Network Construction Methods
1 Introduction
2 Materials and Methods
2.1 Participants and Data Preprocessing
2.2 Construction of the BFN
2.3 Evaluation of Brain Age Prediction
2.4 Evaluation Metrics
3 Results and Discussion
3.1 The Brain Functional Network (BFN)
3.2 Prediction Performance
3.3 Consensus Features and Discriminative Brain Regions
4 Conclusion and Limitation
References
Transfer Dataset in Image Segmentation Use Case
1 Introduction
2 Transfer Dataset: Our Approach
2.1 Our Preliminary Dataset
2.2 The Change of Task and Data Augmentation
3 Experiments: Settings and Results
3.1 Initial Test of Many Networks
3.2 More Precise Tests of the Best Architecture
4 Conclusions
References
Neural Network Models
A Gaussian Process-Based Incremental Neural Network for Online Regression
1 Introduction
2 Proposed Method
2.1 Overview of SOINN
2.2 Gaussian Process and Bandwidth Matrix Optimization
2.3 Posterior Approximation
2.4 Threshold Region Determination
2.5 Property Discussion
2.6 Algorithm: A Gaussian Process-Based Incremental Neural Network for Online Regression
3 Experimental Results
3.1 Synthetic Data
3.2 Real-World Data
4 Conclusion
References
Analysis on the Boltzmann Machine with Random Input Drifts in Activation Function
1 Introduction
2 Background
3 Main Result
4 Simulations
4.1 Use of Theorem 1
4.2 Use of Theorem 3
5 Conclusion
References
Are Deep Neural Architectures Losing Information? Invertibility is Indispensable
1 Introduction
2 Preliminaries
2.1 Residual Blocks
2.2 Flow-Based Generative Models
2.3 Mutual Information
3 Conditions When Deep Architectures Lose Information
4 Flow-Based Image Restoration Models
4.1 Architecture Overview
4.2 Encoder and Decoder
4.3 Invertible Local Spatial Feature Extraction
4.4 Steps of Flow
4.5 Loss Function
5 Experiments
5.1 Experimental Settings
5.2 Comparison on Image Restoration Tasks
6 Discussion and Future Work
7 Conclusion
References
Automatic Dropout for Deep Neural Networks
1 Introduction
2 Methods
2.1 Automatic Dropout
2.2 Adaptive Dropout Based on Clustering
3 Evaluation
3.1 Experiments
3.2 Results
4 Conclusion
References
Bayesian Randomly Wired Neural Network with Variational Inference for Image Recognition
1 Introduction
2 Background
2.1 Variational Inference
2.2 Bayes by Backprop
2.3 Randomly Wired Neural Networks
3 Methods
3.1 Validation Accuracy and Uncertainty Estimation
3.2 Activation Function
3.3 Network Architecture
4 Experiments
5 Conclusion
References
Brain-Inspired Framework for Image Classification with a New Unsupervised Matching Pursuit Encoding
1 Introduction
2 Methodology
2.1 UMP Encoding
2.2 Unsupervised Kernel Learning
2.3 Neuron Model
2.4 Multi-spike Learning
3 Experiments
3.1 Dataset
3.2 Experimental Settings
3.3 Effect of the Number of Encoding Neurons
3.4 Comparison with Different Temporal-Based Encoding
3.5 Comparison Between Different Learning Methods
3.6 Performance Comparison with Other SNNs
4 Conclusion
References
Estimating Conditional Density of Missing Values Using Deep Gaussian Mixture Model
1 Introduction
2 Density Model for Missing Data
2.1 Gaussian Mixture Model for High Dimensional Data
2.2 Conditional Gaussian Density
2.3 Deep Conditional Gaussian Density for Missing Data
3 Experiments
3.1 Gray-Scale Images
3.2 CelebA Dataset
3.3 Parameters of Conditional Density
4 Conclusion and Future Work
References
Environmentally-Friendly Metrics for Evaluating the Performance of Deep Learning Models and Systems
Abstract
1 Introduction
2 Related Work
3 The Proposed Deep Learning Metrics
4 Experimental Setup and Results
5 Conclusions and Future Work
References
Hybrid Deep Shallow Network for Assessment of Depression Using Electroencephalogram Signals
1 Introduction
2 Hybrid Depression Diagnostic Framework
2.1 CNN-Gated Recurrent Units
2.2 1DCNN-Long Short Term Memory
3 Experiment
3.1 Parameters
3.2 Results and Discussion
4 Conclusion
References
Iterative Imputation of Missing Data Using Auto-Encoder Dynamics
1 Introduction
2 Auto-Encoder Dynamics
3 Imputation Method
4 Experiments
5 Conclusion and Future Work
References
Multi-objective Evolution for Deep Neural Network Architecture Search
1 Introduction
2 Related Work
3 Deep and Convolutional Neural Networks
4 Multi-objective Evolution for Deep Neural Networks
4.1 NSGA-2 and NSGA-3 Algorithms
4.2 Individual Encoding
4.3 Genetic Operators
5 Experiments
6 Conclusion
References
Neural Architecture Search for Extreme Multi-label Text Classification
1 Introduction
2 Related Work
3 Framework and Model
3.1 Architecture Search Phase
3.2 Text Embedding, Attention and Classification Modules
3.3 Analysis of Operations Importance
4 Experimental Results
4.1 Datasets and Evaluation Metrics
4.2 Architecture Search Evaluation
4.3 Performance Evaluation
5 Summary and Outlook
References
Non-linear ICA Based on Cramer-Wold Metric
1 Introduction
2 Related Work
3 Independence Measure by Cramer-Wold Distance
4 Algorithm
5 Experiments
5.1 Comparison with ANICA
5.2 Comparison on Image Dataset
6 Conclusions
References
Oblique Random Forests on Residual Network Features
1 Introduction
2 Convolutional Neural Networks and Random Forests
2.1 Fully Convolutional Networks
2.2 Residual Networks
2.3 Random Forests
2.4 Oblique Random Forests
3 Proposed Solution
4 Experiments
4.1 Datasets
4.2 Experiment Setup
4.3 Results
4.4 Performance of Our Methods
5 Conclusion
References
P2ExNet: Patch-Based Prototype Explanation Network
1 Introduction
2 Related Work
2.1 Post-hoc
2.2 Intrinsic
2.3 Limitations of Existing Methods
3 P2ExNet: The Proposed Approach
3.1 Motivation: An Understandable Reasoning Behavior
3.2 Architecture
3.3 Mathematical Background
3.4 Training Process
4 Datasets
5 Experiments
5.1 P2ExNet: Instance-Based Evaluation
5.2 P2ExNet: Evaluation as a Classifier
5.3 P2ExNet: Sanity Check
5.4 Comparison with Existing Prototype-Based Approaches
6 Conclusion
References
Prediction of Taxi Demand Based on CNN-BiLSTM-Attention Neural Network
1 Introduction
2 Models
2.1 CNN-BiLSTM-Attention Model
2.2 Model Components
3 Experiment
3.1 Dataset
3.2 Preprocessing
3.3 Model Hyperparameters
3.4 Evaluation Metrics
4 Results
4.1 Spatio-temporal Features
4.2 Prediction Results
5 Conclusion and Future Work
References
Pruning Long Short Term Memory Networks and Convolutional Neural Networks for Music Emotion Recognition
1 Introduction
2 Method
2.1 Data Set
2.2 Data Pre-processing
2.3 Network Details
3 Results and Discussion
3.1 LSTM Sequence Length
3.2 Classification Results
3.3 Effects of Network Pruning
4 Conclusion and Future Works
References
Unsupervised Multi-layer Spiking Convolutional Neural Network Using Layer-Wise Sparse Coding
1 Introduction
2 Network Architecture
2.1 SAILnet
2.2 Bilateral Filter
2.3 Convolution and Pooling
2.4 Dropout
2.5 Fully Connected SNN
3 Experiment and Results
3.1 Datasets Description
3.2 Experiments
4 Comparative Analysis
5 Conclusion
References
VAEPP: Variational Autoencoder with a Pull-Back Prior
1 Introduction
2 Background
2.1 VAEs and Learnable Priors
2.2 GANs and Wasserstein Distance
3 Pull-Back Prior
3.1 Intuition of Pull-Back Prior
3.2 How to Obtain D and G
3.3 How to Determine
3.4 The Upper-Bound of Z
4 Training and Sampling
4.1 2-Step Training for VAEPP
4.2 1-Step Training for VAEPP
4.3 Sampling from VAEPP
5 Experiments
5.1 Log-Likelihood Evaluation
5.2 Quality of Sampling
6 Conclusion
7 Derivation of Pull-Back Prior
References
Why Do Deep Neural Networks with Skip Connections and Concatenated Hidden Representations Work?
1 Introduction
2 Related Work
3 Preliminaries
3.1 Plain Network (PlainNet)
3.2 Concatened Network of Hidden Representations (ConcNeXt)
4 Theoretical Study of DNNs with Skip Connections and Concatenated Hidden Represenations
4.1 PlainNet (PlainNet)
4.2 Concatened Network of Aggregated Hidden Representations (ConcNeXt)
5 Experiments
5.1 Experimental Settings
5.2 Results and Discussion
5.3 Main Insights into Models Concatenated Hidden Representations
6 Conclusion
References
Recommender Systems
AMBR: Boosting the Performance of Personalized Recommendation via Learning from Multi-behavior Data
1 Introduction
2 Related Work
3 Design of AMBR Algorithm
3.1 Framework Overview
3.2 Attention Based Representation Learning
3.3 Multi-behavior Matching Function Learning
3.4 Model Learning
4 Experiments
4.1 Experimental Settings
4.2 Performance Comparison
4.3 Ablation Study
4.4 Layer Settings
5 Conclusion
References
Asymmetric Pairwise Preference Learning for Heterogeneous One-Class Collaborative Filtering
1 Introduction
2 Related Work
2.1 One-Class Collaborative Filtering
2.2 Heterogeneous One-Class Collaborative Filtering
3 Our Solution
3.1 Problem Definition
3.2 A Generic Preference Assumption
3.3 Weighting Strategy
3.4 Like-Mined User-Group Construction
3.5 Objective Function and Algorithm
4 Experiments
4.1 Datasets and Evaluation Metrics
4.2 Baselines and Configurations
4.3 Results
5 Conclusions and Future Work
References
DPR-Geo: A POI Recommendation Model Using Deep Neural Network and Geographical Influence
1 Introduction
2 Related Work
3 DPR-Geo: POI Recommendation Using Deep Neural Network and Geographical Influence
3.1 Problem Formulation
3.2 Dimension Reduction Network
3.3 Union Network
3.4 Loss Function
3.5 Geographical Influence
3.6 POI Recommendation
4 Experiments
4.1 Datasets
4.2 Evaluation Metrics
4.3 Optimal
4.4 Performance Comparison
5 Conclusion
References
Feature Aware and Bilinear Feature Equal Interaction Network for Click-Through Rate Prediction
1 Introduction
2 Feature Aware and Bilinear Feature Equal Interaction Network
2.1 Sparse Input and Embedding Layer
2.2 SE-ResNet Layer
2.3 Bilinear Feature Equal Interaction Layer
2.4 Concatenation Layer
2.5 DNN Network Layer
2.6 Output Layer
3 Experiments
3.1 Dataset
3.2 Experiments Setting
3.3 Model Comparison
3.4 Ablation Study
4 Conclusion
References
GFEN: Graph Feature Extract Network for Click-Through Rate Prediction
1 Introduction
2 Related Work
2.1 Factorization Machine
2.2 Feature Generation by Convolutional Neural Networks
3 Our Method
3.1 Basic Framework
3.2 Propagation Process in Ripple Network
3.3 Graph Feature Extract Network
3.4 Objective Function
4 Experiments
4.1 Datasets and Metrics
4.2 Baselines Methods
4.3 Results and Discussion
5 Conclusion
References
JUST-BPR: Identify Implicit Friends with Jump and Stay for Social Recommendation
1 Introduction
2 Discover Implicit Friends with Jump and Stay for Recommendation
2.1 Preliminaries
2.2 Generating Node Corpora with Jump and Stay
2.3 Node Embedding to Discover Implicit Friends
2.4 JUST-BPR: BPR with Jump and Stay
3 Experiments
3.1 Experimental Setup
3.2 Recommendation Performance
3.3 Parameter Effect Analysis
3.4 Implicit Friends Identified by Jump and Stay vs. Meta-Paths
4 Related Work
5 Conclusion
References
Leveraging Knowledge Context Information to Enhance Personalized Recommendation
1 Introduction
2 Related Work
3 Task Formulation
4 Our Proposed Model
4.1 Model Overview
4.2 Input and Embedding Layer
4.3 Knowledge Context Generating Layer
4.4 Knowledge Fuison Layer
4.5 Prediction Layer
4.6 Learning Algorithm
5 Experiment and Evaluation
5.1 Experimental Settings
5.2 Performance Comparison (RQ1)
5.3 Ablation Analysis (RQ2)
5.4 Sensitivity Analysis of Hyper-parameters (RQ3)
6 Conclusion
References
LHRM: A LBS Based Heterogeneous Relations Model for User Cold Start Recommendation in Online Travel Platform
1 Introduction
2 Related Work
3 The Proposed Approach
3.1 Problem Statement
3.2 Notations
3.3 Geohash Algorithm
3.4 LBS Based Heterogeneous Relations Model
4 Experiment
4.1 Compared Methods
4.2 Implementation Details
4.3 Datasets
4.4 Results
5 Conclusion
References
Match4Rec: A Novel Recommendation Algorithm Based on Bidirectional Encoder Representation with the Matching Task
1 Introduction
2 Related Works
3 Methodology
3.1 Model Architecture
3.2 Embedding Layer
3.3 Transformer Layer
3.4 Output Layer
3.5 Model Learning
4 Experiments and Discussions
4.1 Datasets and Baselines
4.2 Evaluation Metrics
4.3 Recommendation Performance
4.4 Ablation Study
4.5 Space and Time Complexity Analysis
5 Conclusion
References
Multi-level Feature Extraction in Time-Weighted Graphical Session-Based Recommendation
1 Introduction
2 Related Work
3 Method
3.1 Notations
3.2 Build a Session Graph
3.3 Graph Neural Network Layer
3.4 Feature Extraction Module (FEM)
3.5 Prediction Layer
3.6 Objective Function
4 Experiments
4.1 Datasets
4.2 Evaluation Metrics
4.3 Parameter Setup
4.4 Baseline Methods
4.5 Comparison with Baseline Methods
4.6 Ablation Analysis
4.7 The Sensitivity of Hyper-parameters
5 Conclusion
References
Time Series Analysis
3ETS+RD-LSTM: A New Hybrid Model for Electrical Energy Consumption Forecasting
1 Introduction
2 Forecasting Model
2.1 Architecture and Features
2.2 Time Series Processing
2.3 Residual Delated LSTM
3 Results
4 Conclusion
References
A Deep Time Series Forecasting Method Integrated with Local-Context Sensitive Features
1 Introduction
2 Related Work
2.1 Time Series Forecasting
2.2 Series Significance Representation
3 Framework
3.1 Optimal Sampling for Contexts Generation
3.2 Integration of Context Significance Features
3.3 Local-Context Sensitive Convolution
4 Experiments
4.1 Experimental Setup
4.2 Forecasting Performance Evaluation
5 Conclusion
References
Benchmarking Adversarial Attacks and Defenses for Time-Series Data
1 Introduction
2 Related Work
3 Method
3.1 Threat Model
3.2 Adversarially Robust Models (using Adversarial Defenses)
3.3 Robust Evaluation (using Adversarial Attacks)
3.4 Dataset
4 Results
4.1 Quantifying the Impact of Denoising Operators
4.2 Sensitivity to Regularization Hyperparameters
4.3 Attacked Examples
5 Conclusion
References
Correlation-Aware Change-Point Detection via Graph Neural Networks
1 Introduction
2 Method for CPD
3 Correlation-Aware Dynamics Model
3.1 Correlation Encoder
3.2 Dynamics Decoder
4 Experiment with Physics Simulations
4.1 Particle-Spring Change-Point Dataset
4.2 Evaluation Metric and Baselines
4.3 Main Results
4.4 Change-Point Type Classification
5 Experiments with Physical Activity Monitoring
6 Conclusion
References
DPAST-RNN: A Dual-Phase Attention-Based Recurrent Neural Network Using Spatiotemporal LSTMs for Time Series Prediction
1 Introduction
2 Model
2.1 Encoder
2.2 Decoder
2.3 Training Procedure
3 Experiments
3.1 Data Acquisition
3.2 Datasets and Setup
3.3 Parameter Settings
3.4 Results: Time Series Prediction
4 Conclusion and Future Work
References
ForecastNet: A Time-Variant Deep Feed-Forward Neural Network Architecture for Multi-step-Ahead Time-Series Forecasting
1 Introduction
2 Motivations and Related Work
3 ForecastNet Architecture
4 ForecastNet Properties
4.1 Time-Variance
4.2 Interleaved Outputs
4.3 Memory Requirements
5 Methods and Datasets
5.1 Datasets
5.2 Models
5.3 Training and Testing
6 Results and Discussion
6.1 Time-Invariance Demonstration
6.2 Model Comparison Error Results
6.3 Model Comparison Box-Whisker Plots
7 Summary and Conclusion
References
Memetic Genetic Algorithms for Time Series Compression by Piecewise Linear Approximation
1 Introduction
2 Preliminaries
3 The Piecewise Linear Approximation Genetic Algorithm
4 Experimental Evaluation
4.1 Data Sets
4.2 Evaluation of Hybridization
4.3 Comparison with Other Evolutionary Algorithms
4.4 Comparison with Bellman
5 Conclusions
References
Sensor Drift Compensation Using Robust Classification Method
1 Introduction
2 Dataset
3 Feature Extraction Algorithm
4 LSTM
5 Results
6 Conclusion
References
SpringNet: Transformer and Spring DTW for Time Series Forecasting
1 Introduction
2 Case Study: Solar Power Forecasting
2.1 Data Sets
2.2 Problem Statement
3 Background
3.1 Transformer
3.2 LogSparse Transformer
3.3 Spring Algorithm
4 Proposed Approach: SpringNet
4.1 Motivation and Novelty
4.2 Model Architecture
4.3 Batch Spring Attention
5 Experimental Setup
6 Results and Discussion
7 Conclusions
References
U-Sleep: A Deep Neural Network for Automated Detection of Sleep Arousals Using Multiple PSGs
1 Introduction
2 Related Work
3 Methodology
3.1 Problem Definition
3.2 Model Architecture
4 Experimental Results and Discussion
4.1 Data Description
4.2 Pre-processing
4.3 Evaluation Metric
4.4 Hyperparameter Settings
4.5 Ablation Study
4.6 Performance Comparison
4.7 Discussion
5 Conclusions
References
Author Index

System requirements

Save as PDF Copy link into clipboard

Schweitzer Fachinformationen

Neural Information Processing

Description

More details

Other editions

Additional editions

Content

System requirements