IEEE Transactions on Neural Networks and Learning Systems, Volume 37, Issue 2, February 2026

2026-02-16 · Source: Computational Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Computer Vision, Robotics & Autonomous Systems · Depth: Expert, short

Summary

The February 2026 issue of IEEE Transactions on Neural Networks and Learning Systems presents 35 articles covering diverse advancements in neural networks and machine learning. Key contributions include a comprehensive survey and taxonomy of Vision Mamba models (pages 505-525), and a new paradigm for efficient and robust model training through Hard Sample Mining (pages 526-546). Medical imaging sees a diffusion model with fuzzy evidence-driven dynamic uncertainty fusion for segmentation (FEU-Diff, pages 547-561), while reinforcement learning is applied to boundary-optimized control of flexible manipulators (pages 562-574) and output consensus in multiagent systems (pages 575-588). Other notable topics include all-to-all connected oscillator Ising machines for associative memory (pages 589-602), coupled tensor decomposition for compact network representation (pages 617-631), and federated learning studies on IoT-edge devices (pages 753-765) and with multi-level prototype-based contrastive learning (FedMPS, pages 781-794). The issue also features work on lightweight reparameterizable integral neural networks for mobile applications (pages 809-821) and a mathematical certification for positivity conditions in neural networks for trustworthy AI (pages 981-996).

Key takeaway

For research scientists and engineers developing robust and efficient AI systems, this collection offers critical insights into emerging techniques. You should investigate Hard Sample Mining to enhance model resilience and consider Vision Mamba architectures for advanced computer vision tasks. Additionally, explore the mathematical certification for positivity conditions to build more trustworthy AI models, particularly in safety-critical applications.

Key insights

This issue advances neural network theory and application across diverse domains, from medical imaging to robotics and trustworthy AI.

Principles

• Hard sample mining improves model robustness.
• Fuzzy logic enhances diffusion model uncertainty fusion.
• Tensor decomposition enables compact network representation.

Method

FEU-Diff integrates fuzzy evidence and dynamic uncertainty fusion into a diffusion model for medical image segmentation. Hard Sample Mining identifies and prioritizes difficult examples during training to improve model efficiency and robustness.

In practice

• Apply Hard Sample Mining for robust model training.
• Utilize FEU-Diff for medical image segmentation.
• Explore Vision Mamba for computer vision tasks.

Topics

• Vision Mamba
• Federated Learning
• Diffusion Models
• Reinforcement Learning
• Hard Sample Mining

Best for: Computer Vision Engineer, Research Scientist, AI Researcher, AI Scientist, AI Student

Related on AIssential

• IEEE Transactions on Neural Networks and Learning Systems, Volume 37, Issue 5, May 2026 — The volume showcases diverse advancements in neural networks, from foundational surveys to specialized applications and novel architectures.
• Reload-Mamba: Hierarchical Anti-Dilution State-Space Modeling for Multi-Class Semantic Segmentation — Reload-Mamba enhances Mamba models for segmentation by preventing detail loss via boundary-aware and uncertainty-driven state restoration.
• ConvVitMamba: Efficient Multiscale Convolution, Transformer, and Mamba-Based Sequence modelling for Hyperspectral Image Classification — ConvVitMamba efficiently classifies hyperspectral images by combining multiscale CNNs, Vision Transformers, and Mamba-inspired sequence mixing.
• Identifying Information from Observations with Uncertainty and Novelty — Identification and sample complexity are unified across computational, statistical, and PAC-Bayes learning frameworks.
• Coarse-to-fine Hierarchical Architecture with Sequential Mamba for Brain Reconstruction — CHASMBrain's dual-stream Mamba and coarse-to-fine fMRI encoding architecture achieves high predictive accuracy and reveals visual cortex functional specialization.
• IEEE Transactions on Artificial Intelligence, Volume 7, Issue 4, April 2026 — The IEEE Transactions on AI presents diverse research spanning reinforcement learning, LLMs, robotics, and federated learning.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Computational Intelligence.