Neural Information Processing : 30th International Conference, ICONIP 2023, Changsha, China, November 20–23, 2023, Proceedings, Part IV / edited by Biao Luo, Long Cheng, Zheng-Guang Wu, Hongyi Li, Chaojie Li.

Format:

Book

Contributor:

Luo, Biao, editor.

Series:

Lecture Notes in Computer Science, 1611-3349 ; 14450

Language:

English

Subjects (All):

Pattern recognition systems.

Data mining.

Machine learning.

Social sciences--Data processing.

Social sciences.

Automated Pattern Recognition.

Data Mining and Knowledge Discovery.

Machine Learning.

Computer Application in Social and Behavioral Sciences.

Local Subjects:

Automated Pattern Recognition.

Data Mining and Knowledge Discovery.

Machine Learning.

Computer Application in Social and Behavioral Sciences.

Physical Description:

1 online resource (594 pages)

Edition:

1st ed. 2024.

Place of Publication:

Singapore : Springer Nature Singapore : Imprint: Springer, 2024.

Summary:

The six-volume set LNCS 14447 until 14452 constitutes the refereed proceedings of the 30th International Conference on Neural Information Processing, ICONIP 2023, held in Changsha, China, in November 2023. The 652 papers presented in the proceedings set were carefully reviewed and selected from 1274 submissions. They focus on theory and algorithms, cognitive neurosciences; human centred computing; applications in neuroscience, neural networks, deep learning, and related fields. .

Contents:

Intro

Preface

Organization

Contents - Part IV

Human Centred Computing

Cross-Modal Method Based on Self-Attention Neural Networks for Drug-Target Prediction

1 Instructions

2 Materials and Approaches

2.1 Benchmark Datasets

2.2 Implementation Process of SANN-DTI

2.3 Adjustment of Parameters

2.4 Evaluation Metrics

3 Experimental Results

3.1 Compared with Baseline Models

3.2 Impact of Each Component on Predicted Performance

4 Case Study

5 Conclusion

References

GRF-GMM: A Trajectory Optimization Framework for Obstacle Avoidance in Learning from Demonstration

1 Introduction

2 Problem Statement

3 Method

3.1 Gaussian Mixture Model/Gaussian Mixture Regression

3.2 Optimization Algorithm: GRF-GMM

4 Simulations and Experiments

4.1 2D Handwriting Letter Task

4.2 Experiment

4.3 Comparisons

5 Conclusions

SLG-NET: Subgraph Neural Network with Local-Global Braingraph Feature Extraction Modules and a Novel Subgraph Generation Algorithm for Automated Identification of Major Depressive Disorder

2 Related Work

2.1 Construction of Braingraph

3.1 Sub-braingraph Sampling and Encoding

3.2 Sub-braingraph Selection and Sub-braingraph's Node Selection by LFE Module

3.3 Sub-braingraph Sketching by GFE Module and Classification

4 Experiments

4.1 Dataset and Parameters Setting

4.2 Overall Evaluation

4.3 S-BFS, LFE, and GFE Modules Analysis

CrowdNav-HERO: Pedestrian Trajectory Prediction Based Crowded Navigation with Human-Environment-Robot Ternary Fusion

2.1 Socially Aware Crowded Navigation

2.2 Simulator for Crowded Navigation

3 Problem Formulation

4 HRO Ternary Fusion Simulator

4.1 Simulator Setting.

4.2 Static Environment Construction and Collision Avoidance

4.3 Crowd Interaction Optimization

5 A Crowded Navigation Framework with HERO Ternary Feature Fusion

5.1 Spatial-Temporal Pedestrian Trajectory Prediction

5.2 Dual-Channel Value Estimation Network

6 Experiments

6.1 Experimental Settings

6.2 Quantitative Evaluations for Crowded Navigation

6.3 Quantitative Evaluation of Impact of Environment on Navigation

6.4 Quantitative Evaluations on Real Pedestrian Dataset

7 Conclusion

Modeling User's Neutral Feedback in Conversational Recommendation

3 Problem Definition

4 Proposed Methods

4.1 Representation Learning

4.2 Action Decision

4.3 Selection Strategies

4.4 Update and Deduction

5 Experiments

5.1 DataSet

5.2 Experimental Settings

5.3 Performance Comparison of NFCR with Existing Models (RQ1)

5.4 Ablation Studies (RQ2)

5.5 Case Study on Neutral Feedback (RQ3)

6 Conclusions

A Domain Knowledge-Based Semi-supervised Pancreas Segmentation Approach

2.1 Semi-supervised Medical Image Segmentation

2.2 Domain Knowledge

3 Methodology

3.1 Loss Function

4 Experiments and Results

4.1 Datasets

4.2 Implementation Details

4.3 Ablation Study

4.4 Comparison Study

Soybean Genome Clustering Using Quantum-Based Fuzzy C-Means Algorithm

2 Preliminaries

2.1 Fuzzy C-Means

2.2 Quantum Computing Concept

3 Proposed Work

3.1 Dataset Preparation

3.2 Quantum Fuzzy C-Means (QFCM) Clustering Approach

4 Experiment and Result

4.1 Experimental Environment

4.2 Datasets Description

4.3 Performance Evaluation

4.4 Results and Discussion

References.

DAMFormer: Enhancing Polyp Segmentation Through Dual Attention Mechanism

2.1 Polyp Segmentation

2.2 Attention Mechanism

3 Proposed Method

3.1 Transformer Encoder

3.2 ConvBlock

3.3 Enhanced Dual Attention Module

3.4 Channel-Wise Scaling

3.5 Effective Feature Fusion

BIN: A Bio-Signature Identification Network for Interpretable Liver Cancer Microvascular Invasion Prediction Based on Multi-modal MRIs

2 Related Works

2.1 MVI Prediction Models Based on MRIs

2.2 MVI Interpretable Deep Models

3 The Proposed Multi-modal Fusion Based BIN Method

4 Experiment and Analysis

4.1 Performance Comparisons

4.2 Qualitative Experiment

Human-to-Human Interaction Detection

3 HID Task

3.1 Problem Definition

3.2 Evaluation Metrics

3.3 The AVA-Interaction Dataset

4 SaMFormer

4.1 Visual Feature Extractor

4.2 The Split Stage

4.3 The Merging Stage

4.4 Training and Inference

5.1 Main Results on AVA-I

5.2 Ablation Study

5.3 Qualitative Results

5.4 Evaluation on BIT and UT

6 Conclusion

Reconstructing Challenging Hand Posture from Multi-modal Input

3 Capture

4 Skeleton-Shape Alignment

5 Data Evaluation and Applications

6 Conclusions and Future Work

A Compliant Elbow Exoskeleton with an SEA at Interaction Port

2 Mechanical Design

2.1 Exoskeleton Design

2.2 SEA Analysis

3 SEA Modeling

3.1 NARMAX Model

3.2 T-S Fuzzy Model

3.3 LSTM Model

3.4 Model Training

3.5 Model Validation

4 Exoskeleton Flexible Control

Applications.

Differential Fault Analysis Against AES Based on a Hybrid Fault Model

2 DFA on AES State

2.1 Proposed Fault Model

2.2 The Analysis of Cracking AES

2.3 The Process of Cracking AES

3 Experimental Results and Comparisons

4 Conclusions

Towards Undetectable Adversarial Examples: A Steganographic Perspective

2.1 Adversarial Attack

2.2 Embedding Suitability Map

3 Proposed Scheme

3.1 Motivation

3.2 Embedding Suitability Map-Weighted Attack

3.3 Combination with CAM

4 Experimental Results

4.1 Attack Ability

4.2 Undetectability

4.3 Undetectability-Attack Ability Tradeoff

4.4 Visual Quality

On Efficient Federated Learning for Aerial Remote Sensing Image Classification: A Filter Pruning Approach

2.1 Efficient Federated Learning

2.2 Filter Pruning

3.1 System Model

3.2 Cross-All-Layers Importance Measure for Pruning

3.3 CALIM-FL Work Process

4.1 Experimental Settings

4.2 Result Discussion

ASGNet: Adaptive Semantic Gate Networks for Log-Based Anomaly Diagnosis

3 The Proposed Model

3.1 Task Description

3.2 Definition of Terms

3.3 Log Statistics Information Representation

3.4 Log Deep Semantic Representation

3.5 Adaptive Semantic Threshold Mechanism

4 Experimental Setup

4.1 Dataset and Hyper-parameters

4.2 Training and Hyperparameters

5 Experimental Results

5.1 Model Comparisons (RQ1)

5.2 Ablation Study (RQ2)

5.3 Parameter Sensitivity (RQ3)

Propheter: Prophetic Teacher Guided Long-Tailed Distribution Learning

3 Proposed Method.

3.1 Prophetic Teacher Learning

3.2 Propheter-Guided Long-Tailed Classification

4.1 Datasets and Implementation Details

4.2 Experimental Results

Sequential Transformer for End-to-End Person Search

2 Method

2.1 SeqTR Architecture

2.2 re-ID Transformer

2.3 Training and Inference

3 Experiments

3.1 Datasets and Settings

3.2 Implementation Details

3.3 Comparison to the State-of-the-arts

3.4 Ablation Study

4 Conclusion

Multi-scale Structural Asymmetric Convolution for Wireframe Parsing

2 Methodology

2.1 Overall Network Architecture

2.2 Customized Backbone

2.3 Geometry Proposal Network

3.1 Datasets and Metrics

3.3 Ablation Study

3.4 Comparison with Other Methods

S3ACH: Semi-Supervised Semantic Adaptive Cross-Modal Hashing

3 The Proposed Method

3.1 Notation and Problem Formulation

3.2 S3ACHMethod

3.3 Optimization

3.4 Hash Function Learning

3.5 Time Cost Analysis

4.2 Compared Baselines and Evaluation Metrics

4.3 Implementation Details

4.4 Results

4.5 Ablation Experiments

4.6 Parameter Sensitivity Analysis

4.7 Convergence Analysis

Intelligent UAV Swarm Planning Based on Undirected Graph Model

2 Methods

2.1 Improved MINCO Algorithm

2.2 UAV Cluster Modeling

3 Constraints in Cost Functions

3.1 Smoothness Penalty

3.2 Total Time Penalty

3.3 Collision Penalty

3.4 Cluster Formation Penalty

3.5 Penalty for Collisions Between Unmanned Aerial Vehicles

3.6 Dynamic Feasibility Penalty.

3.7 Penalty for Uniform Distribution of Constraint Points.

Notes:

Includes bibliographical references and index.

ISBN:

981-9980-70-4