Telecomunicaciones, información y comunicación

Feature Disentangling and Combination Implemented by Spin–Orbit Torque Magnetic Tunnel Junctions

15/01/2026 - Spin–orbit torque magnetic tunnel junctions (SOT‐MTJs) enable efficient feature disentangling and integration in image data. A proposed algorithm leverages SOT‐MTJs as true random number...

Human‐Machine Mutual Trust Based Shared Control Framework for Intelligent Vehicles

15/01/2026 - This work introduces a bidirectional‐trust‐driven shared control framework for human‐machine co‐driving. The method models human‐to‐machine trust from intention discrepancies and Bayesian...

Upsampling DINOv2 Features for Unsupervised Vision Tasks and Weakly Supervised Materials Segmentation

15/01/2026 - Feature from recent image foundation models (DINOv2) are useful for vision tasks (segmentation, object localization) with little or no human input. Once upsampled, they can be used for weakly...

A Generalizable Transformer Framework for Gene Regulatory Network Inference from Single‐Cell Transcriptomes

15/01/2026 - FTGRN introduces an LLM‐enhanced framework for gene regulatory network inference through a two‐stage workflow. It combines a Transformer‐based model, pretrained on GPT‐4 derived gene...

Training robots with natural and lightweight human feedback

09/01/2026 - Generalist robot models promise broad applicability across domains but currently require extensive expert demonstrations for task specialization, which is a costly and impractical barrier for...

Programmable‐Stiffness Tensegrity Continuum Robot for Adaptive Multicurvature On‐Orbit Assembly

06/01/2026 - A bioinspired tensegrity‐based continuum robot (BTCR) with programmable stiffness enables adaptive multicurvature morphing via coordinated intra‐ and intermodule regulation. A unified...

Roadmap on Artificial Intelligence‐Augmented Additive Manufacturing

06/01/2026 - This Roadmap outlines the transformative role of artificial intelligence‐augmented additive manufacturing, highlighting advances in design, monitoring, and product development. By integrating tools...

Biomarker‐Agnostic Detection of Ovarian Cancer from Blood Plasma Using a Machine Learning‐Driven Electronic Nose

06/01/2026 - This study introduces a biomarker‐agnostic diagnostic strategy for ovarian cancer, utilizing a machine learning‐enhanced electronic nose to analyze volatile organic compound signatures from blood...

A Novel Milli‐Scale Magnetic Robot Exploiting Rotation for Controlled Magnetic Particles Release

06/01/2026 - Delivering magnetic particles can become a game changer in minimally invasive medicine. To cope with this challenge, a magnetically actuated milli‐scale carrier leveraging rotation to perform...

ControlIt: A Universal Framework for Translational, Adaptive, and Online Brain–Computer Interfaces

06/01/2026 - Brain–computer interfaces (BCIs) lack a unified platform that works across signals and algorithms. ControlIt, an open‐source, modular ROS2‐based BCI framework supporting electroencephalography...

Slip‐Adaptive Neural Control of Gecko‐Inspired Adhesive Robots

06/01/2026 - This study introduces a neural adhesion controller to improve the stability of gecko‐inspired climbing robots. By integrating an echo state network and a multilayer perceptron, the system utilizes...

gnSPADE: Incorporating Gene Network Structures Enhances Reference‐Free Deconvolution in Spatial Transcriptomics

06/01/2026 - gnSPADE integrates gene‐network structures into a probabilistic topic modeling framework to achieve reference‐free cell‐type deconvolution in spatial transcriptomics. By embedding gene...

Predicting Materials Thermodynamics Enabled by Large Language Model‐Driven Dataset Building and Machine Learning

06/01/2026 - Illustration of text data mining of rare earth mineral thermodynamic parameters with the large language model‐powered LMExt. A dataset is built with mined thermodynamic properties. Subsequently, a...

AI Guided Protein Design for Next‐Generation Autogenic Engineered Living Materials

06/01/2026 - Autogenic engineered living materials (ELMs) integrate biology and materials science to create self‐regenerating and self‐healing materials. This perspective highlights emerging strategies in...

Metalearning‐Driven Inverse Optimization for Precision Microstructure Fabrication in Digital Light Processing Three‐Dimensional Printing

06/01/2026 - Metalearning‐based inverse optimization enables precise microscale three‐dimensional printing using a DLP system. Distorted structures from conventional printing are analyzed via neural network...

Real‐Time Sampling‐Based Model Predictive Control Based on Reverse Kullback–Leibler Divergence and Its Adaptive Acceleration

06/01/2026 - This study presents a new sampling‐based model predictive control minimizing reverse Kullback‐Leibler divergence to quickly find a local optimum. In addition, a modified Nesterov's acceleration...

Real‐Time, Label‐Free Classification of Cell Death Pathways via Holotomography‐Based Deep Learning Framework

06/01/2026 - A 3D holotomography system coupled with a deep learning model distinguishes how cells die—apoptosis, necroptosis or necrosis—without any fluorescent labels. Training on refractive index maps of...

Soft Artificial Ciliary Brush with Integrated Haptic Feedback for Efficient Airway Mucus Cleaning

06/01/2026 - A fundamental mucus cleaning mechanism is reported, which utilizes both artificial cilia mechanical sweeping and suction‐based mucus cleaning. A novel haptic feedback interface for the artificial...

Polymerase Chain Reaction. Perturbation Theory and Machine Learning Artificial Intelligence‐Experimental Microbiome Analysis: Applications to Ancient DNA and Tree Soil Metagenomics Cases of Study

06/01/2026 - The polymerase chain reaction (PCR).Perturbation Theory and Machine Learning framework integrates perturbation theory and machine learning to classify genetic sequences, distinguishing ancient DNA...

Adaptive Autonomy in Microrobot Motion Control via Deep Reinforcement Learning and Path Planning Synergy

06/01/2026 - This study introduces a data‐driven framework that combines deep reinforcement learning with classical path planning to achieve adaptive microrobot navigation. By training a surrogate neural...

Maneuverable Multilegged Locomotion through Anisotropically Arranged Soft Backbones in a Single‐Actuator Modular Miniature Robot

06/01/2026 - A modular eight‐legged robot exploits anisotropically oriented soft I‐beam backbones to transmit vibration from a single unbalanced‐mass actuator, producing frequency‐dependent multimodal...