Machine-Learning-Enhanced Printed Vertical Magnetoresistive Sensors for Transparent, Flexible, Multimodal Interactive Magnetoelectronics.

To meet the increasingly stringent demands of next-generation electronic systems, magnetoresistive sensors are required to simultaneously deliver environmental compatibility, advanced functionality, and enhanced intelligence. Here, we demonstrate a synergistic strategy spanning device, algorithm, and system levels to address these challenges in a unified manner. By rationally designing functional inks, fully printable magnetoresistive sensors are realized through additive manufacturing, substantially reducing energy consumption and material waste during fabrication. Introducing magnetic-field guidance during printing enables vertical alignment of functional nanowires, resulting in an out-of-plane sensor architecture. This configuration not only reduces nanowire surface coverage, imparting exceptional optical transparency, but also suppresses the adverse influence of inter-nanowire junctions on electrical percolation, thereby enhancing mechanical robustness. Beyond materials and device engineering, the integration of machine-learning algorithms and system-level optimization extends sensor operation beyond conventional threshold-based mechanisms, enabling robust multi-pattern recognition capabilities. Notably, this functionality is achieved using a single sensing element without relying on sensor matrices or additional electronic components, thus preserving the intrinsic transparency and mechanical flexibility of the system. Leveraging the synergistic combination of these achievements, the proposed sensors offer an eco-responsible platform for next-generation imperceptible and intelligent magnetic sensing.