Search

Non-Newtonian fluids can be used for the protection of flexible laminates. Understanding the coupling between the flow of the protecting fluid and the deformation of the protected solids is necessary in order to optimise this functionality. We present a scaling analysis of the problem based on a single coupling variable, the effective width of a squeeze flow between flat rigid plates, and predict that impact protection for laminates is optimised by using shear-thinning, and not shear-thickening, fluids. The prediction is verified experimentally by measuring the velocity and pressure in impact experiments. Our scaling analysis should be generically applicable for non-Newtonian fluid-solid interactions in diverse applications., Comment: 7 pages, 5 figures with SI of 2 pages, 3 figures

The ability to manipulate rheological and electrical properties of colloidal carbon black gels makes them attractive in composites for energy applications such as batteries and fuel cells, where they conduct electricity and prevent sedimentation of `granular' active components. While it is commonly assumed that granular fillers have a simple additive effect on the composite properties, new phenomena can emerge unexpectedly, with some composites exhibiting a unique rheological bi-stability between solid-like and liquid-like states. Here we report such bi-stability in suspensions of non-Brownian graphite and colloidal carbon black in oil, a model system to mimic composite suspensions for energy applications. Steady shear below a critical stress elicits a transition to a persistent mechanically weak and poorly conducting state, which must be `rejuvenated' using high-stress shear to recover a stronger, high-conductivity state. Our findings highlight the highly tunable nature of binary granular/gel composite suspensions, and present new possibilities for optimizing mixing and processing conditions for Li-ion battery slurries.

The rapid digitization of real-world data offers an unprecedented opportunity for optimizing healthcare delivery and accelerating biomedical discovery. In practice, however, such data is most abundantly available in unstructured forms, such as clinical notes in electronic medical records (EMRs), and it is generally plagued by confounders. In this paper, we present TRIALSCOPE, a unifying framework for distilling real-world evidence from population-level observational data. TRIALSCOPE leverages biomedical language models to structure clinical text at scale, employs advanced probabilistic modeling for denoising and imputation, and incorporates state-of-the-art causal inference techniques to combat common confounders. Using clinical trial specification as generic representation, TRIALSCOPE provides a turn-key solution to generate and reason with clinical hypotheses using observational data. In extensive experiments and analyses on a large-scale real-world dataset with over one million cancer patients from a large US healthcare network, we show that TRIALSCOPE can produce high-quality structuring of real-world data and generates comparable results to marquee cancer trials. In addition to facilitating in-silicon clinical trial design and optimization, TRIALSCOPE may be used to empower synthetic controls, pragmatic trials, post-market surveillance, as well as support fine-grained patient-like-me reasoning in precision diagnosis and treatment., Comment: 6 Figures, 22 Pages, 3 Tables

The ability to manipulate spins in magnetic materials is essential in designing spintronics devices. One method for magnetic switching is through strain. In VO2 on TiO2 thin films, while VO2 remains rutile across the metal-insulator transition, the in-plane lattice area expands going from low temperature insulating phase to high temperature conducting phase. In a VO2/TbFeCo bilayer, the expansion of the VO2 lattice area exerts tension on the amorphous TbFeCo layer. Through the strain effect, magnetic properties, including the magnetic anisotropy and magnetization, of TbFeCo can be changed. In this work, the changes in magnetic properties of TbFeCo on VO2/TiO2(011) are demonstrated using anomalous Hall effect measurements. Across the metal-insulator transition, TbFeCo loses perpendicular magnetic anisotropy, and the magnetization in TbFeCo turns from out-of-plane to in-plane. Using atomistic simulations, we confirm these tunable magnetic properties originating from the metal-insulator transition of VO2. This study provides the groundwork for controlling magnetic properties through a phase transition., Comment: To be published in Nanomaterials

Temporal Knowledge Graph Completion (TKGC) under the extrapolation setting aims to predict the missing entity from a fact in the future, posing a challenge that aligns more closely with real-world prediction problems. Existing research mostly encodes entities and relations using sequential graph neural networks applied to recent snapshots. However, these approaches tend to overlook the ability to skip irrelevant snapshots according to entity-related relations in the query and disregard the importance of explicit temporal information. To address this, we propose our model, Re-Temp (Relation-Aware Temporal Representation Learning), which leverages explicit temporal embedding as input and incorporates skip information flow after each timestamp to skip unnecessary information for prediction. Additionally, we introduce a two-phase forward propagation method to prevent information leakage. Through the evaluation on six TKGC (extrapolation) datasets, we demonstrate that our model outperforms all eight recent state-of-the-art models by a significant margin., Comment: Findings of EMNLP 2023

The recent success of general-domain large language models (LLMs) has significantly changed the natural language processing paradigm towards a unified foundation model across domains and applications. In this paper, we focus on assessing the performance of GPT-4, the most capable LLM so far, on the text-based applications for radiology reports, comparing against state-of-the-art (SOTA) radiology-specific models. Exploring various prompting strategies, we evaluated GPT-4 on a diverse range of common radiology tasks and we found GPT-4 either outperforms or is on par with current SOTA radiology models. With zero-shot prompting, GPT-4 already obtains substantial gains ($\approx$ 10% absolute improvement) over radiology models in temporal sentence similarity classification (accuracy) and natural language inference ($F_1$). For tasks that require learning dataset-specific style or schema (e.g. findings summarisation), GPT-4 improves with example-based prompting and matches supervised SOTA. Our extensive error analysis with a board-certified radiologist shows GPT-4 has a sufficient level of radiology knowledge with only occasional errors in complex context that require nuanced domain knowledge. For findings summarisation, GPT-4 outputs are found to be overall comparable with existing manually-written impressions., Comment: EMNLP 2023 main

The Kitaev chiral spin liquid has captured widespread interest in recent decades because of its intrinsic non-Abelian excitations, yet the experimental realization is challenging. Here we propose to realize and detect Kitaev chiral spin liquid in a deformed honeycomb array of Rydberg atoms. Through a novel laser-assisted dipole-dipole interaction mechanism to generate both effective hopping and pairing terms for hard-core bosons, together with van der Waals interaction, we achieve the Kitaev spin liquid model with high precision. The gapped non-Abelian spin liquid phase is then obtained. Further, we propose to probe the chiral Majorana edge modes by light Bragg scattering and by imagining their chiral motion, in which only spin degree of freedom needs to be measured. Our work broadens the range of exotic quantum many-body phases that can be realized and detected in atomic systems, and makes an important step toward manipulating non-Abelian anyons.

The segmentation of cells and neurites in microscopy images of neuronal networks provides valuable quantitative information about neuron growth and neuronal differentiation, including the number of cells, neurites, neurite length and neurite orientation. This information is essential for assessing the development of neuronal networks in response to extracellular stimuli, which is useful for studying neuronal structures, for example, the study of neurodegenerative diseases and pharmaceuticals. However, automatic and accurate analysis of neuronal structures from phase contrast images has remained challenging. To address this, we have developed NeuroQuantify, an open-source software that uses deep learning to efficiently and quickly segment cells and neurites in phase contrast microscopy images. NeuroQuantify offers several key features: (i) automatic detection of cells and neurites; (ii) post-processing of the images for the quantitative neurite length measurement based on segmentation of phase contrast microscopy images, and (iii) identification of neurite orientations. The user-friendly NeuroQuantify software can be installed and freely downloaded from GitHub https://github.com/StanleyZ0528/neural-image-segmentation.

Rapid progress has been made in instruction-learning for image editing with natural-language instruction, as exemplified by InstructPix2Pix. In biomedicine, such methods can be applied to counterfactual image generation, which helps differentiate causal structure from spurious correlation and facilitate robust image interpretation for disease progression modeling. However, generic image-editing models are ill-suited for the biomedical domain, and counterfactual biomedical image generation is largely underexplored. In this paper, we present BiomedJourney, a novel method for counterfactual biomedical image generation by instruction-learning from multimodal patient journeys. Given a patient with two biomedical images taken at different time points, we use GPT-4 to process the corresponding imaging reports and generate a natural language description of disease progression. The resulting triples (prior image, progression description, new image) are then used to train a latent diffusion model for counterfactual biomedical image generation. Given the relative scarcity of image time series data, we introduce a two-stage curriculum that first pretrains the denoising network using the much more abundant single image-report pairs (with dummy prior image), and then continues training using the counterfactual triples. Experiments using the standard MIMIC-CXR dataset demonstrate the promise of our method. In a comprehensive battery of tests on counterfactual medical image generation, BiomedJourney substantially outperforms prior state-of-the-art methods in instruction image editing and medical image generation such as InstructPix2Pix and RoentGen. To facilitate future study in counterfactual medical generation, we plan to release our instruction-learning code and pretrained models., Comment: Project page & demo: https://aka.ms/biomedjourney

Single photon detection is important for a wide range of low-light applications, including quantum information processing, spectroscopy, and light detection and ranging (LiDAR). A key challenge in these applications has been to integrate single-photon detection capability into photonic circuits for the realization of complex photonic microsystems. Short-wavelength ($\lambda$ < 1.1 $\mu$m) integrated photonics platforms that use silicon (Si) as photodetectors offer the opportunity to achieve single-photon avalanche diodes (SPADs) that operate at or near room temperature. Here, we report the first waveguide-coupled Si SPAD. The device is monolithically integrated in a Si photonic platform and operates in the visible spectrum. The device exhibited a single photon detection efficiency of > 6% for wavelengths of 488 nm and 532 nm with an excess voltage less than 20% of the breakdown voltage. The dark count rate was below 100 kHz at room temperature, with the possibility of improving by approximately 35% by reducing the temperature to -5$^{\circ}$C.

Optimal Transport is a useful metric to compare probability distributions and to compute a pairing given a ground cost. Its entropic regularization variant (eOT) is crucial to have fast algorithms and reflect fuzzy/noisy matchings. This work focuses on Inverse Optimal Transport (iOT), the problem of inferring the ground cost from samples drawn from a coupling that solves an eOT problem. It is a relevant problem that can be used to infer unobserved/missing links, and to obtain meaningful information about the structure of the ground cost yielding the pairing. On one side, iOT benefits from convexity, but on the other side, being ill-posed, it requires regularization to handle the sampling noise. This work presents an in-depth theoretical study of the l1 regularization to model for instance Euclidean costs with sparse interactions between features. Specifically, we derive a sufficient condition for the robust recovery of the sparsity of the ground cost that can be seen as a far reaching generalization of the Lasso's celebrated Irrepresentability Condition. To provide additional insight into this condition, we work out in detail the Gaussian case. We show that as the entropic penalty varies, the iOT problem interpolates between a graphical Lasso and a classical Lasso, thereby establishing a connection between iOT and graph estimation, an important problem in ML.

There are profound connections between neutrino physics and nuclear experiments. Exceptionally precise measurements of single and double beta-decay spectra illuminate the scale and nature of neutrino mass and may finally answer the question of whether neutrinos are their own antimatter counterparts. Neutrino-nucleus scattering underpins oscillation experiments and probes nuclear structure, neutrinos offer a rare vantage point into collapsing stars and nuclear fission reactors, and techniques pioneered in neutrino nuclear-physics experiments are advancing quantum-sensing technologies. In this article, we review current and planned efforts at the intersection of neutrino and nuclear experiments., Comment: 22 pages, 4 figures, 1 table. Submitted as a contribution to "The liminal position of Nuclear Physics: from hadrons to neutron stars" issue of Philosophical Transactions A. v1: submitted version; v2: accepted version

The use of optimal transport (OT) distances, and in particular entropic-regularised OT distances, is an increasingly popular evaluation metric in many areas of machine learning and data science. Their use has largely been driven by the availability of efficient algorithms such as the Sinkhorn algorithm. One of the drawbacks of the Sinkhorn algorithm for large-scale data processing is that it is a two-phase method, where one first draws a large stream of data from the probability distributions, before applying the Sinkhorn algorithm to the discrete probability measures. More recently, there have been several works developing stochastic versions of Sinkhorn that directly handle continuous streams of data. In this work, we revisit the recently introduced online Sinkhorn algorithm of [Mensch and Peyr\'e, 2020]. Our contributions are twofold: We improve the convergence analysis for the online Sinkhorn algorithm, the new rate that we obtain is faster than the previous rate under certain parameter choices. We also present numerical results to verify the sharpness of our result. Secondly, we propose the compressed online Sinkhorn algorithm which combines measure compression techniques with the online Sinkhorn algorithm. We provide numerical experiments to show practical numerical gains, as well as theoretical guarantees on the efficiency of our approach.

Due to the low flux of exoEarths, long exposure times are required to spectrally characterize them. During these long exposures, the contrast in the dark hole will degrade as the the optical system drifts from its initial DH state. To prevent such contrast drift, a wavefront sensing and control (WFSC) algorithm running in parallel to the science acquisition can stabilize the contrast. However, pairwise probing (PWP) cannot be reused to efficiently stabilize the contrast since it relies on strong temporal modulation of the intensity in the image plane, which would interrupt the science acquisition. The use of small amplitude probes has been demonstrated but requires multiple measurements from each science sub-band to converge. Conversely, spectral linear dark field control (LDFC) takes advantage of the linear relationship between the change in intensity of the post-coronagraph out-of-band image and small changes in wavefront in the science band to preserve the DH region during science exposures. In this paper, we show experimental results that demonstrate spectral LDFC stabilizes the contrast to levels of a few $10^{-9}$ on a Lyot coronagraph testbed which is housed in a vacuum chamber. Promising results show that spectral LDFC is able to correct for disturbances that degrade the contrast by more than 100$\times$. To our knowledge, this is the first experimental demonstration of spectral LDFC and the first demonstration of spatial or spectral LDFC on a vacuum coronagraph testbed and at contrast levels less than $10^{-8}$., Comment: 14 pages, 10 figures, Techniques and Instrumentation for Detection of Exoplanets XI

Two bounded linear operators $A$ and $B$ are parallel with respect to a norm $\|\cdot\|$ if $\|A+\mu B\| = \|A\| + \|B\|$ for some scalar $\mu$ with $|\mu| = 1$. Characterization is obtained for bijective linear maps sending parallel bounded linear operators to parallel bounded linear operators with respect to the Ky-Fan $k$-norms., Comment: 20 pages

Future space telescope coronagraph instruments hinge on the integration of high-performance masks and precise wavefront sensing and control techniques to create dark holes essential for exoplanet detection. Recent advancements in wavefront control algorithms might exhibit differing performance depending on the coronagraph used. This research investigates three model-free and model-based algorithms in conjunction with either a vector vortex coronagraph or a scalar vortex coronagraph under identical laboratory conditions: pairwise probing with electric field conjugation, the self-coherent camera with electric field conjugation, and implicit electric field conjugation. We present experimental results in narrowband and broadband light from the In-Air Coronagraph Testbed at the Jet Propulsion Laboratory. We find that model-free dark hole digging methods achieve comparable broadband contrasts to model-based methods, and highlight the calibration costs of model-free methods compared to model-based approaches. This study also reports the first time that electric field conjugation with the self-coherent camera has been applied for simultaneous multi-subband correction with a field stop. This study compares the advantages and disadvantages of each of these wavefront sensing and control algorithms with respect to their potential for future space telescopes.

The enclosed data release consists of a subset of the calibration data from the Majorana Demonstrator experiment. Each Majorana event is accompanied by raw Germanium detector waveforms, pulse shape discrimination cuts, and calibrated final energies, all shared in an HDF5 file format along with relevant metadata. This release is specifically designed to support the training and testing of Artificial Intelligence (AI) and Machine Learning (ML) algorithms upon our data. This document is structured as follows. Section I provides an overview of the dataset's content and format; Section II outlines the location of this dataset and the method for accessing it; Section III presents the NPML Machine Learning Challenge associated with this dataset; Section IV contains a disclaimer from the Majorana collaboration regarding the use of this dataset; Appendix A contains technical details of this data release. Please direct questions about the material provided within this release to liaobo77@ucsd.edu (A. Li)., Comment: DataPlanet Access: https://dataplanet.ucsd.edu/dataset.xhtml?persistentId=perma:83.ucsddata/UQWQAV

Extracting meaningful drug-related information chunks, such as adverse drug events (ADE), is crucial for preventing morbidity and saving many lives. Most ADEs are reported via an unstructured conversation with the medical context, so applying a general entity recognition approach is not sufficient enough. In this paper, we propose a new multi-aspect cross-integration framework for drug entity/event detection by capturing and aligning different context/language/knowledge properties from drug-related documents. We first construct multi-aspect encoders to describe semantic, syntactic, and medical document contextual information by conducting those slot tagging tasks, main drug entity/event detection, part-of-speech tagging, and general medical named entity recognition. Then, each encoder conducts cross-integration with other contextual information in three ways: the key-value cross, attention cross, and feedforward cross, so the multi-encoders are integrated in depth. Our model outperforms all SOTA on two widely used tasks, flat entity detection and discontinuous event extraction., Comment: Accepted at CIKM 2023

An innovative method is proposed to construct a quantile dependence system for inflation and money growth. By considering all quantiles and leveraging a novel notion of quantile sensitivity, the method allows the assessment of changes in the entire distribution of a variable of interest in response to a perturbation in another variable's quantile. The construction of this relationship is demonstrated through a system of linear quantile regressions. Then, the proposed framework is exploited to examine the distributional effects of money growth on the distributions of inflation and its disaggregate measures in the United States and the Euro area. The empirical analysis uncovers significant impacts of the upper quantile of the money growth distribution on the distribution of inflation and its disaggregate measures. Conversely, the lower and median quantiles of the money growth distribution are found to have a negligible influence. Finally, this distributional impact exhibits variation over time in both the United States and the Euro area.

We propose a novel image registration method based on implicit neural representations that addresses the challenging problem of registering a pair of brain images with similar anatomical structures, but where one image contains additional features or artifacts that are not present in the other image. To demonstrate its effectiveness, we use 2D microscopy $\textit{in situ}$ hybridization gene expression images of the marmoset brain. Accurately quantifying gene expression requires image registration to a brain template, which is difficult due to the diversity of patterns causing variations in visible anatomical brain structures. Our approach uses implicit networks in combination with an image exclusion loss to jointly perform the registration and decompose the image into a support and residual image. The support image aligns well with the template, while the residual image captures individual image characteristics that diverge from the template. In experiments, our method provided excellent results and outperformed other registration techniques., Comment: 11 pages

Despite the growing number of confident binary black hole coalescences observed through gravitational waves so far, the astrophysical origin of these binaries remains uncertain. Orbital eccentricity is one of the clearest tracers of binary formation channels. Identifying binary eccentricity, however, remains challenging due to the limited availability of gravitational waveforms that include effects of eccentricity. Here, we present observational results for a waveform-independent search sensitive to eccentric black hole coalescences, covering the third observing run (O3) of the LIGO and Virgo detectors. We identified no new high-significance candidates beyond those that were already identified with searches focusing on quasi-circular binaries. We determine the sensitivity of our search to high-mass (total mass $M>70$ $M_\odot$) binaries covering eccentricities up to 0.3 at 15 Hz orbital frequency, and use this to compare model predictions to search results. Assuming all detections are indeed quasi-circular, for our fiducial population model, we place an upper limit for the merger rate density of high-mass binaries with eccentricities $0 < e \leq 0.3$ at $0.33$ Gpc$^{-3}$ yr$^{-1}$ at 90\% confidence level., Comment: 24 pages, 5 figures

Large language models (LLMs) have demonstrated remarkable generalizability, such as understanding arbitrary entities and relations. Instruction tuning has proven effective for distilling LLMs into more cost-efficient models such as Alpaca and Vicuna. Yet such student models still trail the original LLMs by large margins in downstream applications. In this paper, we explore targeted distillation with mission-focused instruction tuning to train student models that can excel in a broad application class such as open information extraction. Using named entity recognition (NER) for case study, we show how ChatGPT can be distilled into much smaller UniversalNER models for open NER. For evaluation, we assemble the largest NER benchmark to date, comprising 43 datasets across 9 diverse domains such as biomedicine, programming, social media, law, finance. Without using any direct supervision, UniversalNER attains remarkable NER accuracy across tens of thousands of entity types, outperforming general instruction-tuned models such as Alpaca and Vicuna by over 30 absolute F1 points in average. With a tiny fraction of parameters, UniversalNER not only acquires ChatGPT's capability in recognizing arbitrary entity types, but also outperforms its NER accuracy by 7-9 absolute F1 points in average. Remarkably, UniversalNER even outperforms by a large margin state-of-the-art multi-task instruction-tuned systems such as InstructUIE, which uses supervised NER examples. We also conduct thorough ablation studies to assess the impact of various components in our distillation approach. We release the distillation recipe, data, and UniversalNER models to facilitate future research on targeted distillation., Comment: Accepted at ICLR 2024. Project page: https://universal-ner.github.io/

Clinical trial matching is a key process in health delivery and discovery. In practice, it is plagued by overwhelming unstructured data and unscalable manual processing. In this paper, we conduct a systematic study on scaling clinical trial matching using large language models (LLMs), with oncology as the focus area. Our study is grounded in a clinical trial matching system currently in test deployment at a large U.S. health network. Initial findings are promising: out of box, cutting-edge LLMs, such as GPT-4, can already structure elaborate eligibility criteria of clinical trials and extract complex matching logic (e.g., nested AND/OR/NOT). While still far from perfect, LLMs substantially outperform prior strong baselines and may serve as a preliminary solution to help triage patient-trial candidates with humans in the loop. Our study also reveals a few significant growth areas for applying LLMs to end-to-end clinical trial matching, such as context limitation and accuracy, especially in structuring patient information from longitudinal medical records., Comment: 24 pages, 5 figures, accepted at Machine Learning for Healthcare (MLHC) 2023

Document understanding and information extraction include different tasks to understand a document and extract valuable information automatically. Recently, there has been a rising demand for developing document understanding among different domains, including business, law, and medicine, to boost the efficiency of work that is associated with a large number of documents. This workshop aims to bring together researchers and industry developers in the field of document intelligence and understanding diverse document types to boost automatic document processing and understanding techniques. We also released a data challenge on the recently introduced document-level VQA dataset, PDFVQA. The PDFVQA challenge examines the structural and contextual understandings of proposed models on the natural full document level of multiple consecutive document pages by including questions with a sequence of answers extracted from multi-pages of the full document. This task helps to boost the document understanding step from the single-page level to the full document level understanding., Comment: Accepted at CIKM 2023; Orgnised in CIKM

We present 294 pulsars found in GeV data from the Large Area Telescope (LAT) on the Fermi Gamma-ray Space Telescope. Another 33 millisecond pulsars (MSPs) discovered in deep radio searches of LAT sources will likely reveal pulsations once phase-connected rotation ephemerides are achieved. A further dozen optical and/or X-ray binary systems co-located with LAT sources also likely harbor gamma-ray MSPs. This catalog thus reports roughly 340 gamma-ray pulsars and candidates, 10% of all known pulsars, compared to $\leq 11$ known before Fermi. Half of the gamma-ray pulsars are young. Of these, the half that are undetected in radio have a broader Galactic latitude distribution than the young radio-loud pulsars. The others are MSPs, with 6 undetected in radio. Overall, >235 are bright enough above 50 MeV to fit the pulse profile, the energy spectrum, or both. For the common two-peaked profiles, the gamma-ray peak closest to the magnetic pole crossing generally has a softer spectrum. The spectral energy distributions tend to narrow as the spindown power $\dot E$ decreases to its observed minimum near $10^{33}$ erg s$^{-1}$, approaching the shape for synchrotron radiation from monoenergetic electrons. We calculate gamma-ray luminosities when distances are available. Our all-sky gamma-ray sensitivity map is useful for population syntheses. The electronic catalog version provides gamma-ray pulsar ephemerides, properties and fit results to guide and be compared with modeling results., Comment: 142 pages. Accepted by the Astrophysical Journal Supplement

When it comes to storing 3D city models in a database, the implementation of the CityGML data model can be quite demanding and often results in complicated schemas. As an example, 3DCityDB, a widely used solution, depends on a schema having 66 tables, mapping closely the CityGML architecture. In this paper, we propose an alternative (called cjdb) for storing CityGML models efficiently in PostgreSQL with a much simpler table structure and data model design (only 3 tables are necessary). This is achieved by storing the attributes and geometries of the objects directly in JSON. In the case of the geometries we thus adopt the Simple Feature paradigm and we use the structure of CityJSON. We compare our solution against 3DCityDB with large real-world 3D city models, and we find that cjdb has significantly lower demands in storage space (around a factor of 10), allows for faster import/export of data, and has a comparable data retrieval speed with some queries being faster and some slower. The accompanying software (importer and exporter) is available at https://github.com/cityjson/cjdb/ under a permissive open-source license.

Large language models (LLMs), such as GPT-4, have demonstrated remarkable capabilities across a wide range of tasks, including health applications. In this paper, we study how LLMs can be used to scale biomedical knowledge curation. We find that while LLMs already possess decent competency in structuring biomedical text, by distillation into a task-specific student model through self-supervised learning, substantial gains can be attained over out-of-box LLMs, with additional advantages such as cost, efficiency, and white-box model access. We conduct a case study on adverse drug event (ADE) extraction, which is an important area for improving care. On standard ADE extraction evaluation, a GPT-3.5 distilled PubMedBERT model attained comparable accuracy as supervised state-of-the-art models without using any labeled data. Despite being over 1,000 times smaller, the distilled model outperformed its teacher GPT-3.5 by over 6 absolute points in F1 and GPT-4 by over 5 absolute points. Ablation studies on distillation model choice (e.g., PubMedBERT vs BioGPT) and ADE extraction architecture shed light on best practice for biomedical knowledge extraction. Similar gains were attained by distillation for other standard biomedical knowledge extraction tasks such as gene-disease associations and protected health information, further illustrating the promise of this approach.

The hydrodynamic regime of the Stribeck curve giving the friction coefficient $\mu$ as a function of the dimensionless relative sliding speed (the Sommerfeld number, $S$) of two contacting non-conformal surfaces is usually considered trivial, with $\mu \sim S$. We predict that for conformal surfaces contacting over large areas, a combination of independent length scales gives rise to a universal power-law with a non-trivial exponent, $\mu\sim S^{2/3}$, for a thick lubrication film. Deviations as the film thins (decreasing $S$) may superficially resemble the onset of elastohydrodynamic lubrication, but are due to a crossover between hydrodynamic regimes. Our experiments as well as recent measurements of chocolate lubrication confirm these predictions., Comment: 5 pages, 4 figures, RevTeX 4.2; supplemental material as SM.pdf