arXiv subjects

Kevin Brown

Publications and source records attributed to Kevin Brown.

16 recordsLinked to original sources

arXiv2026

Evaluating Forecasting Techniques for Hardware Errors on a Large-scale HPC System

Hardware error logs in high-performance computing (HPC) systems provide early signals of abnormal behavior, yet there remain challenges in effectively forecasting these errors using modern predictive methods. This work investigates the boundaries of applying time series forecasting to HPC hardware error dynamics. We use seven years of production logs from the Theta supercomputer to evaluate the predictive efficacy of classical statistical and deep learning models. Our results show that forecasting effectiveness depends strongly on the temporal structure of the error series: regularly occurring and structurally stable errors can be modeled accurately, particularly by LSTM and Transformer architectures with temporal features, while sparse and burst-dominated errors remain difficult to predict. Rather than proposing a deployment-ready failure prediction framework, this study provides empirical guidance on when forecasting is effective and highlights potential directions for improving forecasting accuracy in HPC hardware error analysis.

cs.LG↗

arXiv2026

The Patient is not a Moving Document: A World Model Training Paradigm for Longitudinal EHR

Large language models (LLMs) trained with next-word-prediction have achieved success as clinical foundation models. Representations from these language backbones yield strong linear probe performance across biomedical tasks, suggesting that patient semantics emerge from next-token prediction at scale. However, this paradigm treats patients as a document to be summarized rather than a dynamical system to be simulated; a patient's trajectory emerges from their state evolving under interventions and time, requiring models that simulate dynamics rather than predict tokens. To address this, we introduce SMB-Structure, a world model for structured EHR that grounds a joint-embedding prediction architecture (JEPA) with next-token prediction (SFT). SFT grounds our model to reconstruct future patient states in token space, while JEPA predicts those futures in latent space from the initial patient representation alone, forcing trajectory dynamics to be encoded before the next state is observed. We validate across two large-scale cohorts: Memorial Sloan Kettering (23,319 oncology patients; 323,000+ patient-years) and INSPECT (19,402 pulmonary embolism patients). Using a linear probe evaluated at multiple points along the disease trajectory, we demonstrate that our training paradigm learns embeddings that capture disease dynamics not recoverable by autoregressive baselines, enabling SMB-Structure to achieve competitive performance on complex tasks characterized by high patient heterogeneity. Model weights are available at https://huggingface.co/standardmodelbio/SMB-v1-1.7B-Structure.

cs.AI↗

arXiv2025

GenVarFormer: Predicting gene expression from long-range mutations in cancer

Distinguishing the rare "driver" mutations that fuel cancer progression from the vast background of "passenger" mutations in the non-coding genome is a fundamental challenge in cancer biology. A primary mechanism that non-coding driver mutations contribute to cancer is by affecting gene expression, potentially from millions of nucleotides away. However, existing predictors of gene expression from mutations are unable to simultaneously handle interactions spanning millions of base pairs, the extreme sparsity of somatic mutations, and generalize to unseen genes. To overcome these limitations, we introduce GenVarFormer (GVF), a novel transformer-based architecture designed to learn mutation representations and their impact on gene expression. GVF efficiently predicts the effect of mutations up to 8 million base pairs away from a gene by only considering mutations and their local DNA context, while omitting the vast intermediate sequence. Using data from 864 breast cancer samples from The Cancer Genome Atlas, we demonstrate that GVF predicts gene expression with 26-fold higher correlation across samples than current models. In addition, GVF is the first model of its kind to generalize to unseen genes and samples simultaneously. Finally, we find that GVF patient embeddings are more informative than ground-truth gene expression for predicting overall patient survival in the most prevalent breast cancer subtype, luminal A. GVF embeddings and gene expression yielded concordance indices of $0.706^{\pm0.136}$ and $0.573^{\pm0.234}$, respectively. Our work establishes a new state-of-the-art for modeling the functional impact of non-coding mutations in cancer and provides a powerful new tool for identifying potential driver events and prognostic biomarkers.

q-bio.GN↗

arXiv2025

Building the EHR Foundation Model via Next Event Prediction

Electronic Health Records (EHRs) contain rich temporal dynamics that conventional encoding approaches fail to adequately capture. While Large Language Models (LLMs) show promise for EHR modeling, they struggle to reason about sequential clinical events and temporal dependencies. We propose Next Event Prediction (NEP), a framework that enhances LLMs' temporal reasoning through autoregressive fine-tuning on clinical event sequences. By reformulating EHRs as timestamped event chains and predicting future medical events, NEP explicitly models disease progression patterns and causal relationships. Extensive evaluations across oncology survival prediction and clinical diagnosis tasks demonstrate NEP's superiority, outperforming specialized EHR models by 4.6% AUROC and general-purpose LLMs by 7.2% C-index in temporal reasoning tasks. Our analyses reveal dual benefits: state-of-the-art prediction accuracy combined with clinically interpretable attention patterns that align with known disease pathways.

cs.AI↗

arXiv2025

Patient-specific Biomolecular Instruction Tuning

Proteomics data is essential to pathogenic understanding of a disease phenotype. In cancer, analysis of molecular signatures enables precision medicine through the identification of biological processes that drive individualized tumor progression, therapeutic resistance, and clinical heterogeneity. Recent advances in multimodal large language models (LLMs) have shown remarkable capacity to integrate and reason across heterogeneous data modalities. However, performing multi-modal language modeling for molecular understanding of patient-specific proteomics remains a significant challenge due to two barriers: (1) the lack of instruction-tuning datasets that enable clinical interpretation from proteomics data, and (2) the absence of language modeling architectures designed to capture the rich heterogeneity of molecular data. In this work, we introduce CPTAC-PROTSTRUCT, the first instruction tuning dataset for molecular understanding of oncology, comprising over 400k open-ended examples derived from individualized proteomic profiles curated from the largest national proteomics cancer study (CPTAC). Additionally, we propose KRONOS (Knowledge Representation of patient Omics Networks in Oncology via Structured tuning), a novel graph-LLM framework that leverages molecular interaction topology with proteomics to learn patient-specific graph representations for enhanced clinical reasoning. We show that KRONOS achieves competitive performance across benchmark clinical tasks, including molecular classification, temporal trajectory modeling, and tumor stage prediction from proteomics data. Ultimately, this approach empowers LLMs to understand patient-level pathogenesis, advancing precision medicine through more accurate diagnosis, prognosis, and treatment stratification.

q-bio.QM↗

arXiv2025

Status of the Proton EDM Experiment (pEDM)

The Proton EDM Experiment (pEDM) is the first direct search for the proton electric dipole moment (EDM) with the aim of being the first experiment to probe the Standard Model (SM) prediction of any particle EDM. Phase-I of pEDM will achieve $10^{-29} e\cdot$cm, improving current indirect limits by four orders of magnitude. This will establish a new standard of precision in nucleon EDM searches and offer a unique sensitivity to better understand the Strong CP problem. The experiment is ideally positioned to explore physics beyond the Standard Model (BSM), with sensitivity to axionic dark matter via the signal of an oscillating proton EDM and across a wide mass range of BSM models from $\mathcal{O}(1\text{GeV})$ to $\mathcal{O}(10^3\text{TeV})$. Utilizing the frozen-spin technique in a highly symmetric storage ring that leverages existing infrastructure at Brookhaven National Laboratory (BNL), pEDM builds upon the technological foundation and experimental expertise of the highly successful Muon $g$$-$$2$ Experiments. With significant R\&D and prototyping already underway, pEDM is preparing a conceptual design report (CDR) to offer a cost-effective, high-impact path to discovering new sources of CP violation and advancing our understanding of fundamental physics. It will play a vital role in complementing the physics goals of the next-generation collider while simultaneously contributing to sustaining particle physics research and training early-career researchers during gaps between major collider operations.

hep-ex↗

arXiv2024

Advancing High Resolution Vision-Language Models in Biomedicine

Multi-modal learning has significantly advanced generative AI, especially in vision-language modeling. Innovations like GPT-4V and open-source projects such as LLaVA have enabled robust conversational agents capable of zero-shot task completions. However, applying these technologies in the biomedical field presents unique challenges. Recent initiatives like LLaVA-Med have started to adapt instruction-tuning for biomedical contexts using large datasets such as PMC-15M. Our research offers three key contributions: (i) we present a new instruct dataset enriched with medical image-text pairs from Claude3-Opus and LLaMA3 70B, (ii) we propose a novel image encoding strategy using hierarchical representations to improve fine-grained biomedical visual comprehension, and (iii) we develop the Llama3-Med model, which achieves state-of-the-art zero-shot performance on biomedical visual question answering benchmarks, with an average performance improvement of over 10% compared to previous methods. These advancements provide more accurate and reliable tools for medical professionals, bridging gaps in current multi-modal conversational assistants and promoting further innovations in medical AI.

cs.CL↗

arXiv2023

Language Models are Few-shot Learners for Prognostic Prediction

Clinical prediction is an essential task in the healthcare industry. However, the recent success of transformers, on which large language models are built, has not been extended to this domain. In this research, we explore the use of transformers and language models in prognostic prediction for immunotherapy using real-world patients' clinical data and molecular profiles. This paper investigates the potential of transformers to improve clinical prediction compared to conventional machine learning approaches and addresses the challenge of few-shot learning in predicting rare disease areas. The study benchmarks the efficacy of baselines and language models on prognostic prediction across multiple cancer types and investigates the impact of different pretrained language models under few-shot regimes. The results demonstrate significant improvements in accuracy and highlight the potential of NLP in clinical research to improve early detection and intervention for different diseases.

cs.CL↗

arXiv2022

Masked Image Modeling Advances 3D Medical Image Analysis

Recently, masked image modeling (MIM) has gained considerable attention due to its capacity to learn from vast amounts of unlabeled data and has been demonstrated to be effective on a wide variety of vision tasks involving natural images. Meanwhile, the potential of self-supervised learning in modeling 3D medical images is anticipated to be immense due to the high quantities of unlabeled images, and the expense and difficulty of quality labels. However, MIM's applicability to medical images remains uncertain. In this paper, we demonstrate that masked image modeling approaches can also advance 3D medical images analysis in addition to natural images. We study how masked image modeling strategies leverage performance from the viewpoints of 3D medical image segmentation as a representative downstream task: i) when compared to naive contrastive learning, masked image modeling approaches accelerate the convergence of supervised training even faster (1.40$\times$) and ultimately produce a higher dice score; ii) predicting raw voxel values with a high masking ratio and a relatively smaller patch size is non-trivial self-supervised pretext-task for medical images modeling; iii) a lightweight decoder or projection head design for reconstruction is powerful for masked image modeling on 3D medical images which speeds up training and reduce cost; iv) finally, we also investigate the effectiveness of MIM methods under different practical scenarios where different image resolutions and labeled data ratios are applied.

cs.CV↗

arXiv2022

The storage ring proton EDM experiment

We describe a proposal to search for an intrinsic electric dipole moment (EDM) of the proton with a sensitivity of \targetsens, based on the vertical rotation of the polarization of a stored proton beam. The New Physics reach is of order $10^~3$TeV mass scale. Observation of the proton EDM provides the best probe of CP-violation in the Higgs sector, at a level of sensitivity that may be inaccessible to electron-EDM experiments. The improvement in the sensitivity to $\theta_{QCD}$, a parameter crucial in axion and axion dark matter physics, is about three orders of magnitude.

hep-ph↗

arXiv2022

Electric dipole moments and the search for new physics

Static electric dipole moments of nondegenerate systems probe mass scales for physics beyond the Standard Model well beyond those reached directly at high energy colliders. Discrimination between different physics models, however, requires complementary searches in atomic-molecular-and-optical, nuclear and particle physics. In this report, we discuss the current status and prospects in the near future for a compelling suite of such experiments, along with developments needed in the encompassing theoretical framework.

hep-ph↗

arXiv2021

Mapping Quantum Circuits to Ions in Storage Ring Quantum Computer Architectures

The mapping of quantum circuits to qubits represented by ion states in a storage ring quantum computer is examined. Serial, parallel and hybrid architectures for mapping such qubits onto a storage ring quantum computer are presented. Parallelism is an important part of the SRQC architecture in terms of multiple ions, windows and lasers. A "predecessor problem" that arises in such architectures is identified and a solution is proposed. Representative numerical sizing and timing calculations are presented. The entire methodology is very general and extends to several implementation options. Open problems are identified.

quant-ph↗

arXiv2020

VerSe: A Vertebrae Labelling and Segmentation Benchmark for Multi-detector CT Images

Vertebral labelling and segmentation are two fundamental tasks in an automated spine processing pipeline. Reliable and accurate processing of spine images is expected to benefit clinical decision-support systems for diagnosis, surgery planning, and population-based analysis on spine and bone health. However, designing automated algorithms for spine processing is challenging predominantly due to considerable variations in anatomy and acquisition protocols and due to a severe shortage of publicly available data. Addressing these limitations, the Large Scale Vertebrae Segmentation Challenge (VerSe) was organised in conjunction with the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) in 2019 and 2020, with a call for algorithms towards labelling and segmentation of vertebrae. Two datasets containing a total of 374 multi-detector CT scans from 355 patients were prepared and 4505 vertebrae have individually been annotated at voxel-level by a human-machine hybrid algorithm (https://osf.io/nqjyw/, https://osf.io/t98fz/). A total of 25 algorithms were benchmarked on these datasets. In this work, we present the the results of this evaluation and further investigate the performance-variation at vertebra-level, scan-level, and at different fields-of-view. We also evaluate the generalisability of the approaches to an implicit domain shift in data by evaluating the top performing algorithms of one challenge iteration on data from the other iteration. The principal takeaway from VerSe: the performance of an algorithm in labelling and segmenting a spine scan hinges on its ability to correctly identify vertebrae in cases of rare anatomical variations. The content and code concerning VerSe can be accessed at: https://github.com/anjany/verse.

cs.CV↗

arXiv2015

Generating Configurable Hardware from Parallel Patterns

In recent years the computing landscape has seen an in- creasing shift towards specialized accelerators. Field pro- grammable gate arrays (FPGAs) are particularly promising as they offer significant performance and energy improvements compared to CPUs for a wide class of applications and are far more flexible than fixed-function ASICs. However, FPGAs are difficult to program. Traditional programming models for reconfigurable logic use low-level hardware description languages like Verilog and VHDL, which have none of the pro- ductivity features of modern software development languages but produce very efficient designs, and low-level software lan- guages like C and OpenCL coupled with high-level synthesis (HLS) tools that typically produce designs that are far less efficient. Functional languages with parallel patterns are a better fit for hardware generation because they both provide high-level abstractions to programmers with little experience in hard- ware design and avoid many of the problems faced when gen- erating hardware from imperative languages. In this paper, we identify two optimizations that are important when using par- allel patterns to generate hardware: tiling and metapipelining. We present a general representation of tiled parallel patterns, and provide rules for automatically tiling patterns and gen- erating metapipelines. We demonstrate experimentally that these optimizations result in speedups up to 40x on a set of benchmarks from the data analytics domain.

cs.DC↗

arXiv2015

The Cornell-BNL FFAG-ERL Test Accelerator: White Paper

The Cornell-BNL FFAG-ERL Test Accelerator (C$\beta$) will comprise the first ever Energy Recovery Linac (ERL) based on a Fixed Field Alternating Gradient (FFAG) lattice. In particular, we plan to use a Non Scaling FFAG (NS-FFAG) lattice that is very compact and thus space- and cost- effective, enabling multiple passes of the electron beam in a single recirculation beam line, using the superconducting RF (SRF) linac multiple times. The FFAG-ERL moves the cost optimized linac and recirculation lattice to a dramatically better optimum. The prime accelerator science motivation for C$\beta$ is proving that the FFAG-ERL concept works. This is an important milestone for the Brookhaven National Laboratory (BNL) plans to build a major Nuclear Physics facility, eRHIC, based on producing 21 GeV electron beams to collide with the RHIC ion beams. A consequence of the C$\beta$ work would be the availability of significantly better, cost-effective, compact CW high-brightness electron beams for a plethora of scientific investigations and applications, such as X-ray sources, dark-matter and dark-energy searches, and industrial high-power Free-Electron Laser (FEL) applications. C$\beta$ brings together the resources and expertise of a large DOE National Laboratory, BNL, and a leading research university, Cornell. C$\beta$ will be built in an existing building at Cornell, for the most part using components that have been developed under previous R&D programs, including a fully commissioned world-leading photoemission electron injector, a large SRF accelerator module, and a high-power beam stop. The only elements that require design and construction from scratch is the FFAG magnet transport lattice. This white paper describes a project that promises to propel high-power, high-brightness electron beam science and applications to an exciting new level.

physics.acc-ph↗

arXiv2015

Sloppiness and Emergent Theories in Physics, Biology, and Beyond

Large scale models of physical phenomena demand the development of new statistical and computational tools in order to be effective. Many such models are `sloppy', i.e., exhibit behavior controlled by a relatively small number of parameter combinations. We review an information theoretic framework for analyzing sloppy models. This formalism is based on the Fisher Information Matrix, which we interpret as a Riemannian metric on a parameterized space of models. Distance in this space is a measure of how distinguishable two models are based on their predictions. Sloppy model manifolds are bounded with a hierarchy of widths and extrinsic curvatures. We show how the manifold boundary approximation can extract the simple, hidden theory from complicated sloppy models. We attribute the success of simple effective models in physics as likewise emerging from complicated processes exhibiting a low effective dimensionality. We discuss the ramifications and consequences of sloppy models for biochemistry and science more generally. We suggest that the reason our complex world is understandable is due to the same fundamental reason: simple theories of macroscopic behavior are hidden inside complicated microscopic processes.

cond-mat.stat-mech↗