Search for: All records

Abstract Automatic differentiation (AD) enables powerful metasurface inverse design but requires extensive theoretical and programming expertise. We present a Model Context Protocol (MCP) assisted framework that allows researchers to conduct inverse design with differentiable solvers through large language models (LLMs). Since LLMs inherently lack knowledge of specialized solvers, our proposed solution provides dynamic access to verified code templates and comprehensive documentation through dedicated servers. The LLM autonomously accesses these resources to generate complete inverse design codes without prescribed coordination rules. Evaluation on the Huygens meta-atom design task with the differentiable TorchRDIT solver shows that while both natural language and structured prompting strategies achieve high success rates, structured prompting significantly outperforms in design quality, workflow efficiency, computational cost, and error reduction. The minimalist server design, using only 5 APIs, demonstrates how MCP makes sophisticated computational tools accessible to researchers without programming expertise, offering a generalizable integration solution for other scientific tasks. 

ABSTRACT Human infections with the protozoanLophomonashave been increasingly reported in the medical literature over the past three decades. Initial reports were based on microscopic identification of the purported pathogen in respiratory specimens. Later, a polymerase chain reaction (PCR) was developed to detectLophomonas blattarum, following which there has been a significant increase in reports. In this minireview, we thoroughly examine the published reports ofLophomonasinfection to evaluate its potential role as a human pathogen. We examined the published images and videos of purportedLophomonas,compared its morphology and motility characteristics with host bronchial ciliated epithelial cells and trueL. blattarumderived from cockroaches, analyzed the published PCR that is being used for its diagnosis, and reviewed the clinical data of patients reported in the English and Chinese literature. From our analysis, we conclude that the images and videos from human specimens do not represent trueLophomonasand are predominantly misidentified ciliated epithelial cells. Additionally, we note that there is insufficient clinical evidence to attribute the cases toLophomonasinfection, as the clinical manifestations are non-specific, possibly caused by other infections and comorbidities, and there is no associated tissue pathology attributable toLophomonas. Finally, our analysis reveals that the published PCR is not specific toLophomonasand can amplify DNA from commensal trichomonads. Based on this thorough review, we emphasize the need for rigorous scientific scrutiny before a microorganism is acknowledged as a novel human pathogen and discuss the potential harms of misdiagnoses for patient care and scientific literature.