Search for: All records

Under a high frequency drive, Josephson junctions demonstrate Shapiro steps of quantized voltage. These are dynamically stabilized states in which the phase across the junction locks to the external drive. We explore the stochastic switching between two symmetric steps at \hbar \omega/2e and − \hbar \omega/2e. Surprisingly, the switching rate exhibits a pronounced nonmonotonicity as a function of temperature, violating the general expectation that transitions should become faster with temperature. We explain this behavior by realizing that the system retains memory of the dynamic state from which it is switching, thereby breaking the conventional simplifying assumptions about separations of timescales. 

Artificial Intelligence (AI) has demonstrated significant potential in healthcare, particularly in disease diagnosis and treatment planning. Recent progress in Medical Large Vision-Language Models (Med-LVLMs) has opened up new possibilities for interactive diagnostic tools. However, these models often suffer from factual hallucination, which can lead to incorrect diagnoses. Fine-tuning and retrieval-augmented generation (RAG) have emerged as methods to address these issues. However, the amount of high-quality data and distribution shifts between training data and deployment data limit the application of fine-tuning methods. Although RAG is lightweight and effective, existing RAG-based approaches are not sufficiently general to different medical domains and can potentially cause misalignment issues, both between modalities and between the model and the ground truth. In this paper, we propose a versatile multimodal RAG system, MMed-RAG, designed to enhance the factuality of Med-LVLMs. Our approach introduces a domain-aware retrieval mechanism, an adaptive retrieved contexts selection, and a provable RAG-based preference fine-tuning strategy. These innovations make the RAG process sufficiently general and reliable, significantly improving alignment when introducing retrieved contexts. Experimental results across five medical datasets (involving radiology, ophthalmology, pathology) on medical VQA and report generation demonstrate that MMed-RAG can achieve an average improvement of 43.8% in factual accuracy in the factual accuracy of Med-LVLMs. 

Abstract We use in situ measurements from the first 19 encounters of Parker Solar Probe and the most recent five encounters of Solar Orbiter to study the evolution of the turbulent sonic Mach numberM_t(the ratio of the amplitude of velocity fluctuations to the sound speed) with radial distance and its relationship to density fluctuations. We focus on the near-Sun region with radial distances ranging from about 11 to 80R_⊙. Our results show that (1) the turbulent sonic Mach numberM_tgradually moves toward larger values as it approaches the Sun, until at least 11R_⊙, whereM_tis much larger than the previously observed value of 0.1 at and above 0.3 au; (2) transonic turbulence withM_t ∼ 1 is observed in situ for the first time and is found mostly near the Alfvén critical surface; (3) Alfvén Mach number of the bulk flowM_Ashows a strong correlation with the plasma beta, indicating that most of the observed sub-Alfvénic intervals correspond to a low-beta plasma; (4) the scaling relation between density fluctuations andM_tgradually changes from a linear scaling at larger radial distances to a quadratic scaling at smaller radial distances; and (5) transonic turbulence is more compressible than subsonic turbulence, with enhanced density fluctuations and slightly flatter spectra than subsonic turbulence. A systematic understanding of compressible turbulence near the Sun is necessary for future solar wind modeling efforts.