arXiv subjects

Mark Nguyen

Publications and source records attributed to Mark Nguyen.

5 recordsLinked to original sources

arXiv2026

Localized efficient in-vacuum loading of $\sim$0.1-10 $\mu$m spherical and plate-like particles into optical traps using a pulled glass capillary

We demonstrate a compact piezoelectric-driven micropipette launcher for localized in-vacuum delivery of nano- and microparticles into optical traps. The launcher has been integrated into multiple optical trapping setups, including a single-beam trap, a non-interfering dual beam trap, and a standing-wave dual beam trap, showcasing the versatility and ease of integration of the setup. Using the micropipette launcher, we have successfully trapped silica spheres of $170\text{ nm}$, $300\text{ nm}$, 3 $\mu\text{m}$ diameter, as well as 6 $\mu\text{m}\times$ 0.2 $\mu\text{m}$ $\beta$-NaYF hexagonal prisms and $\sim 100$ nm diameter high-purity nanodiamonds. We characterize the performance of the device including the peak acceleration, angular distribution of emitted particles, and the dependence on vertical displacement between the pipette tip and optical trap. Trapping efficiency as high as 93\% is achieved.

physics.optics↗

arXiv2025

Automated Triaging and Transfer Learning of Incident Learning Safety Reports Using Large Language Representational Models

PURPOSE: Incident reports are an important tool for safety and quality improvement in healthcare, but manual review is time-consuming and requires subject matter expertise. Here we present a natural language processing (NLP) screening tool to detect high-severity incident reports in radiation oncology across two institutions. METHODS AND MATERIALS: We used two text datasets to train and evaluate our NLP models: 7,094 reports from our institution (Inst.), and 571 from IAEA SAFRON (SF), all of which had severity scores labeled by clinical content experts. We trained and evaluated two types of models: baseline support vector machines (SVM) and BlueBERT which is a large language model pretrained on PubMed abstracts and hospitalized patient data. We assessed for generalizability of our model in two ways. First, we evaluated models trained using Inst.-train on SF-test. Second, we trained a BlueBERT_TRANSFER model that was first fine-tuned on Inst.-train then on SF-train before testing on SF-test set. To further analyze model performance, we also examined a subset of 59 reports from our Inst. dataset, which were manually edited for clarity. RESULTS Classification performance on the Inst. test achieved AUROC 0.82 using SVM and 0.81 using BlueBERT. Without cross-institution transfer learning, performance on the SF test was limited to an AUROC of 0.42 using SVM and 0.56 using BlueBERT. BlueBERT_TRANSFER, which was fine-tuned on both datasets, improved the performance on SF test to AUROC 0.78. Performance of SVM, and BlueBERT_TRANSFER models on the manually curated Inst. reports (AUROC 0.85 and 0.74) was similar to human performance (AUROC 0.81). CONCLUSION: In summary, we successfully developed cross-institution NLP models on incident report text from radiation oncology centers. These models were able to detect high-severity reports similarly to humans on a curated dataset.

cs.CL↗

arXiv2025

A method for optically trapping nanospheres at micron range from a tilted mirror

We propose and experimentally demonstrate a novel optical method for trapping and cooling dielectric nanospheres at (sub)-micron distances from a reflective metallic surface. By translating a tilted mirror towards the focus of a single-beam optical tweezer, the optical trap transitions into an off-axis standing-wave configuration due to interference between the incident and reflected beams. Stable potential minima emerge within a finite overlap region close to the surface, with their number, shape, and distance from the surface tunable via the incidence angle, waist, and polarization of the incoming beam. This configuration enables deterministic selection of trapping sites as the system transitions from the single-beam trap to the off-axis standing wave trap. We validate this approach using a $170$ nm diameter silica sphere in a single-beam trap with a $1.5$ $\mu$m waist and transitioning it into the second or first potential minimum of the standing wave trap, located $1.61$ $\mu$m or $0.55$ $\mu$m from the surface, respectively. The experimental results align well with our theoretical model, supported by numerical simulations of the Langevin equations of motion. Additionally, we perform parametric feedback cooling of all three motional degrees of freedom in a high-vacuum environment. This method provides a robust platform for ultra-sensitive scanning surface force sensing at micron distances from a reflective surface in high vacuum and may open new pathways for short-range gravity or Casimir effect measurements.

physics.optics↗

arXiv2023

Semi-automated extraction of research topics and trends from NCI funding in radiological sciences from 2000-2020

Investigators, funders, and the public desire knowledge on topics and trends in publicly funded research but current efforts in manual categorization are limited in scale and understanding. We developed a semi-automated approach to extract and name research topics, and applied this to \$1.9B of NCI funding over 21 years in the radiological sciences to determine micro- and macro-scale research topics and funding trends. Our method relies on sequential clustering of existing biomedical-based word embeddings, naming using subject matter experts, and visualization to discover trends at a macroscopic scale above individual topics. We present results using 15 and 60 cluster topics, where we found that 2D projection of grant embeddings reveals two dominant axes: physics-biology and therapeutic-diagnostic. For our dataset, we found that funding for therapeutics- and physics-based research have outpaced diagnostics- and biology-based research, respectively. We hope these results may (1) give insight to funders on the appropriateness of their funding allocation, (2) assist investigators in contextualizing their work and explore neighboring research domains, and (3) allow the public to review where their tax dollars are being allocated.

cs.CL↗

arXiv2022

The dynamics of viruslike capsid assembly and disassembly

Cowpea chlorotic mottle virus (CCMV) is a widely used model for virus replication studies. A major challenge lies in distinguishing between the roles of the interaction between coat proteins and that between the coat proteins and the viral RNA in assembly and disassembly processes. Here, we report on the spontaneous and reversible size conversion of the empty capsids of a CCMV capsid protein functionalized with a hydrophobic elastin-like polypeptide which occurs following a pH jump. We monitor the concentration of T = 3 and T = 1 capsids as a function of time and show that the time evolution of the conversion from one T number to another is not symmetric: the conversion from T = 1 to T = 3 is a factor of 10 slower than that of T = 3 to T = 1. We explain our experimental findings using a simple model based on classical nucleation theory applied to virus capsids, in which we account for the change in the free protein concentration, as the different types of shells assemble and disassemble by shedding or absorbing single protein subunits. As far as we are aware, this is the first study confirming that both the assembly and disassembly of viruslike shells can be explained through classical nucleation theory, reproducing quantitatively results from time-resolved experiments.

physics.bio-ph↗