Search

Health monitoring systems have revolutionized modern healthcare by enabling the continuous capture of physiological and behavioral data, essential for preventive measures and early health intervention. While integrating this data with Large Language Models (LLMs) has shown promise in delivering interactive health advice, traditional methods like Retrieval-Augmented Generation (RAG) and fine-tuning often fail to fully utilize the complex, multi-dimensional, and temporally relevant data from wearable devices. These conventional approaches typically provide limited actionable and personalized health insights due to their inadequate capacity to dynamically integrate and interpret diverse health data streams. In response, this paper introduces a graph-augmented LLM framework designed to significantly enhance the personalization and clarity of health insights. Utilizing a hierarchical graph structure, the framework captures inter and intra-patient relationships, enriching LLM prompts with dynamic feature importance scores derived from a Random Forest Model. The effectiveness of this approach is demonstrated through a sleep analysis case study involving 20 college students during the COVID-19 lockdown, highlighting the potential of our model to generate actionable and personalized health insights efficiently. We leverage another LLM to evaluate the insights for relevance, comprehensiveness, actionability, and personalization, addressing the critical need for models that process and interpret complex health data effectively. Our findings show that augmenting prompts with our framework yields significant improvements in all 4 criteria. Through our framework, we can elicit well-crafted, more thoughtful responses tailored to a specific patient.

The strong spin-orbit interaction in the organic-inorganic perovskites tied to the incorporation of heavy elements (\textit{e.g.} Pb, I) makes these materials interesting for applications in spintronics. Due to a lack of inversion symmetry associated with distortions of the metal-halide octahedra, the Rashba effect (used \textit{e.g.} in spin field-effect transistors and spin filters) has been predicted to be much larger in these materials than in traditional III-V semiconductors such as GaAs, supported by the recent observation of a near record Rashba spin splitting in CH$_3$NH$_3$PbBr$_3$ using angle-resolved photoemission spectroscopy (ARPES). More experimental studies are needed to confirm and quantify the presence of Rashba effects in the organic-inorganic perovskite family of materials. Here we apply time-resolved circular dichroism techniques to the study of carrier spin dynamics in a 2D perovskite thin film [(BA)$_2$MAPb$_2$I$_7$; BA = CH$_3$(CH$_2$)$_3$NH$_3$, MA = CH$_3$NH$_3$]. Our findings confirm the presence of a Rashba spin splitting via the dominance of precessional spin relaxation induced by the Rashba effective magnetic field. The size of the Rashba spin splitting in our system was extracted from simulations of the measured spin dynamics incorporating LO-phonon and electron-electron scattering, yielding a value of 10 meV at an electron energy of 50 meV above the band gap, representing a 20 times larger value than in GaAs quantum wells.

We have studied the effect of cobalt substitution in $L1_{0}$-Fe$_{1-x}$Co$_{x}$Pt films by means of x-ray magnetic circular dichroism (XMCD) and first-principles calculations. The magnetic moments of Fe ($\sim$2.5 $\mu_{\rm B}$) and Co ($\sim$1.5 $\mu_{\rm B}$) deduced using XMCD were almost unchanged upon Co doping, and the net magnetization decreases with increasing Co content. Calculation also showed that the 3$d$ electrons that have been added by Co substitution occupy only spin-down bands.