The 20-volume set LNCS 15842-15861, together with the 4-volume set LNAI 15862-15865 and the 4-volume set LNBI 15866-15869, constitutes the refereed proceedings of the 21st International Conference on Intelligent Computing, ICIC 2025, held in Ningbo, China, during July 26-29, 2025.

The 1206 papers presented in these proceedings books were carefully reviewed and selected from 4032 submissions. They deal with emerging and challenging topics in artificial intelligence, machine learning, pattern recognition, bioinformatics, and computational biology.

Inhalt

**.- Neural Networks.
** .- Domain Attention and Confidence-Aware Unsupervised Domain Adaptation Network.
.- Enhanced Spatio-Temporal Extended Pattern Diffusion Network for Traffic Flow Forecasting.
.- Research on Digital 3D Reconstruction of Cultural Relics Using Depth Diffusion Gaussian Splatting.
.- Improve Self-supervision Learning by Enhancing Invariant Information.
.- FQuant: Fast Quantization with Adaptive Resolution via the Clustering Algorithm.
.- Solving Seismic Wave Propagation Using an Adaptive Collocation Point[1]Based Physics-Informed Neural Network.
.- Improving Story Visualization via Attribute Encoding and Adaptive Attention.
.- Text Generation Image Model Based on Gated Convolution Attention Generation Adversarial Network.
.- EFF-ViT: A Vision Transformer with Feature Enhancement and Fusion for Fine-Grained Visual Classification.
.- Concept-based Reasoning Explanation for Deep Neural Networks: Drawing on Human Decision-making.
.- OOD Generalization of GNNs through Causal Stabilization Learning.
.- BiGuidedPrompt: Dynamic Bidirectional Guided Multimodal Prompt Learning.
.- Flight Arrival Delay Prediction Based on Bidirectional Temporal Convolutional Networks.
.- Multi-Granularity Filter Pruning for Deep Neural Network Compression.
.- EEGTCT: Electroencephalogram-based Chinese Text Decoding.
.- Decoupled Graph Neural Networks with Hybrid Data Augmentation.
.- Non-invasive Emotion Perception from Gait by Sparse and Spatial-Temporal Excitation based Graph Convolutional Network.
.- Resource-Constrained Scheduling in Containerized Edge Computing Using Graph Transformer-Enhanced DQN.
.- REAMP: A Redundancy Elimination System for AMP-GNN Acceleration.
.- AGD-Net: An Attention-Guided Network for Joint Background Suppression and Defect-Aware Detail Enhancement.
.- Guarding Graph Neural Networks Against Backdoor Attacks-A Training Loss Dynamics Approach.
.- Dual-Resolution Segmentation Network Utilizing Multi-Scale Features for Metal Defect Detection.
.- Robust Multiview Point Cloud Registration with Reliable Sparse Graph and Adaptive Reweighting.
.- Keyphrase Generation Based on the Fusion of Sequence and Word Graph Features.
.- Decoupled Dual-Path Diffusion: Precise Spatial-Semantic Modeling for Human-Object Interaction Generation.
.- Enhancing Virtual Try-On with Text-Image Fusion Guidance.
.- Showing Many Labels in Multi-label Classification Models: An Empirical Study of Adversarial Examples.
.- A Wireless Collaborated Inference Acceleration Framework for Plant Disease Recognition.
.- Unsupervised Learning for Solving the Graph Edit Distance.
.- MCE:One-Shot method to relation extraction based on LLMs.
.- Device Anomaly Sound Detection Based on Unsupervised Adversarial Distillation Domain Adaptation.
.- Multi-Bit Mechanism: Towards Ultra-Low Time Steps for Spiking Neural Networks.
.- FB-SAM: An Effective Learning Framework for First Break Picking Based on the SAM Model with Limited Data.
.- Feature Visualization in 3D Convolutional Neural Networks.
.- Multi-Dimensional Spatiotemporal Modeling for Multimodal Emotion Recognition in Conversations.
.- Multi-Scale Periodic Residual State Space Model for Time Series Forecasting.
.- YOLO-MAOD: An Algorithm for Ground Object Detection in Open-Pit Mine Based on Remote Sensing Images.
.- SimMix: Enhancing Label Consistency in Graph Mixup for Improved Graph Classification.
.- GUIDE: Learnable Deep Contrastive Graph Clustering with Centrality Guidance.
.- LEFF-YOLO: A Lightweight Cherry Tomato Detection YOLOv8 Network with Enhanced Feature Fusion.
.- Hierarchical Text Graph Learning for Inductive Text Classification.
.- Co-GNN: A Co-Optimization Framework for Memory and Computation in Sampling-based GNN Training.
.- FGSMS: Fine-Grained SM Scheduling for Efficient Deep Learning Computing.
.- Spatiotemporal PM10 Concentration Forecasting via Residual Attention Fusion and Mixture-of-Experts Enhanced Graph Neural Network.