The four-volume set LNCS 16068-16071 constitutes the proceedings of the 34th International Conference on Artificial Neural Networks and Machine Learning, ICANN 2025, held in Kaunas, Lithuania, September 9–12, 2025. The 170 full papers and 8 abstracts included in these conference proceedings were carefully reviewed and selected from 375 submissions. The conference strongly values the synergy between theoretical progress and impactful real-world applications, and actively encourages contributions that demonstrate how artificial neural networks are being used to address pressing societal and technological challenges.
.- Epilepsy Prediction based on Intra- and Inter-Channel Feature Mixing. .- Fine-grained Recognition of Arteriovenous Fistula Stenosis Using Blood Flow Sounds: An Animal Model-Based Dataset and a Frequency-Aware Decoupling Network. .- SMART-RetroNet: A Framework for Chemical Retrosynthesis Prediction. .- Few-shot Learning for Syndrome Differentiation with Two Prompts. .- Neural QSLIM for Mesh Autoencoders. .- Evolving Spatially Embedded Recurrent Spiking Neural Networks for Control Tasks. .- Amortizing Personnalization in Virtual Brain Twins. .- MPCCP:A Multi-chain Perception Crime Charge Prediction Method. .- Beran Estimator Kernel Learning using Nearest-Neighbours and its Application to Reliability Analysis. .- Conformalized Causal Learning for Uncertainty-Aware MineralProspectivity Mapping. .- PhysMamba: Synergistic State Space Duality Model for Remote Physiological Measurement. .- Proactive Depot Discovery: A Generative DRL Framework for Adaptive Location-Routing. .- Learning Joint General and Specific Representation with Masked Auto-encoder for Radiology Report Generation. .- Process Adaptive Learning for Visual-Language Navigation. .- Audio-Driven Talking Head Generation with Emotion Based on FLAME Geometry Model. .- Studying the Generalization Behavior of Surrogate Models for Punch-Bending by Generating Plausible Counterfactuals. .- NGAT: A Node-level Graph Attention Network for Long-term Stock Prediction. .- CSM: Corn Instance Segmentation Model Fusing Dilated Residual Networks and Low-Rank Adaptation. .- Sensor-Enhanced PINNs for Contaminant Dispersion Modeling. .- Uniform Representation of Parametric CAD Models for Generative Application. .- Surrogate-Assisted Multi-Objective Design of Complex Multibody Systems. .- KANLoc: WiFi Localization with A Lightweight KAN. .- DualGF: Example-based Path Planning via Dual Gradient Fields. .- Improving physics-informed neural network extrapolation via transfer learning and adaptive activation functions. .- A Spiking Central Pattern Generator Capable of Adaptive Gait Control in Quadruped Locomotion. .- ViSMoE: Visual-Aware Sparse Mixture-of-Experts for Embodied Referring Expression Grounding. .- FDFRL: Credit Card Fraud Detection Based on Federated Reinforcement Learning. .- MENGLAN:Multiscale Enhanced Nonparametric Gas Analyzer with Lightweight Architecture and Networks. .- Targeted trust-based merging of customers’ opinions. .- DA-NeRF: High-Fidelity Talking Face Generation From Speech With Neural Radiance Fields. .- Beyond Reconstruction: A Physics Based Neural Deferred Shader for Photo-realistic Rendering. .- Accurate SDF Reconstruction with Geometric-Differential Regularization and Categorized Sampling Strategy. .- Optimized Supervised Control of Stochastic Timed Discrete Event Systems using Supervisory Control Theory and Reinforcement learning. .- A Classification Algorithm for Bronchiolitis Obliterans in Pediatric CT Images with Extreme Class Imbalance. .- DISEncoder:A Dual-Branch Query Encoder Using Graph Models for Distributed Databases. .- A Subject-Independent Stress Detection Model Based on Temporal Feature Disentanglement.