Pohang University of Science and Technology

Personalized federated recommendation via long-horizon local optimization and regularized knowledge guidance

Abstract Federated Recommendation (FedRec) enables privacy-preserving recommendation by training models across user devices without sharing raw interaction data. However, personalized FedRec still faces a trade-off between personalization and resource efficiency. In this paper, w...

Minimum Partition of Polygons under Width and Cut Constraints

We study the problem of partitioning a polygon into the minimum number of subpolygons using cuts in a fixed set of directions such that each resulting subpolygon satisfies a given width constraint. A polygon satisfies the unit-width constraint for a set of directions if the lengt...

Chemomechanically Adaptive MXene–Si–Ag Dual‐Seed Interlayer Enabling Low‐Pressure Sulfide all‐Solid‐State Batteries

ABSTRACT Sulfide‐based all‐solid‐state batteries (ASSBs) promise high safety and energy density; however, sustaining stable operation under low stack pressure remains challenging due to chemo‐mechanical instability at solid‐solid interfaces. Here, we report a chemomechanically ad...

Building a Physics-Aware AI Ecosystem for Solid-State Hydrogen Storage Materials

Abstract Hydrogen storage remains a central bottleneck for scalable hydrogen energy systems due to the multiscale and coupled nature of the thermodynamics, kinetics, and microstructural evolution of hydrogen storage materials (HSMs). Although artificial intelligence (AI) has acce...

Synergetic Enhancement in Thermoelectric Performance via Localized Magnetic Moments and Lattice Softening

ABSTRACT Magnetism‐assisted transport is achieved in Mn‐containing SnSe magnetic nanocomposites, where superparamagnetic MnSe nanoprecipitates are introduced to modulate both carrier and phonon scattering. Transmission electron microscopy reveals well‐dispersed 5–10 nm MnSe inclu...

Integrating multimodal dot product fusion with conditional generative adversarial network to build a C-P-S-P chain in Mg alloys

Establishing quantitative relationships among composition, process, structure, and property (C-P-S-P) is a fundamental challenge in materials science, yet conventional models relying solely on numerical data fail to capture microstructural spatial heterogeneity, limiting predicti...

Single-vesicle heterogeneity and uptake of extracellular vesicles affected by oncogenic KRAS and BRAF revealed by proteomics and nano-flow cytometry

Extracellular vesicles (EVs) facilitate intercellular communication through the transfer of bioactive molecules that modulate tumor behavior. In this study, we demonstrate that oncogenic KRAS- and BRAF-mutations alter EV composition, modify surface properties, and impact intracel...

Backward Completion Pools for Domain-Specific LLM Agents: Soft Search Modulation with Compiled Relaxations

Production LLM agents serve repeated requests against fixed applications - booking, CRM, project tracking - where goals recur as templates. Current methods allocate test-time search effort using learned signals (process reward models, self-evaluation), which cost extra model call...

Reconfiguration of Localized Ruthenium Surface via Incorporating Single Platinum Atoms for Favorable Hydrogen Oxidation Catalysis

ABSTRACT Despite the fact that only 2% of active sites can satisfy hydrogen oxidation reaction (HOR) activity thanks to fast kinetics, the fuel cell anode is still dependent on catalysts with large amounts of platinum (Pt). Herein, a minimal‐cost ruthenium catalyst bearing ultral...

A sacrificial HfN interlayer enables oxygen-buffered Mo electrodes for sub‑0.4nm equivalent oxide thickness of HfO2/ZrO2 nanolaminates for DRAM capacitors

Continued scaling of dynamic random-access memory (DRAM) capacitors demands simultaneous achievement of high capacitance density and extremely low leakage current at ever-decreasing physical thickness, making interfacial chemistry and defect control as critical as the choice of h...

Fitting-Free Diagnosis of Conduction-Model Breakdown in Laser Powder Bed Fusion

Melt-pool depth governs interlayer bonding and porosity in laser powder bed fusion and underpins part qualification, yet predicting it reliably remains difficult. Fast conduction models reach useful accuracy only after the absorptivity is fitted to the depths they are meant to pr...