检索结果-内蒙古大学图书馆

CANCERLLM: A LARGE LANGUAGE MODEL IN CANCER DOMAIN

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Li, Mingchen Blaes, Anne Johnson, Steven Liu, Hongfang Xu, Hua Zhang, Rui Division of Computational Health Sciences University of Minnesota Twin Cities United States Division of Hematology Oncology and Transplantation University of Minnesota Twin Cities United States Institute for Health Informatics University of Minnesota Twin Cities United States McWilliams School of Biomedical Informatics UTHealth Houston United States Department of Biomedical Informatics and Data Science Yale School of Medicine United States

Medical Large Language Models (LLMs) such as ClinicalCamel 70B, Llama3- OpenBioLLM 70B have demonstrated impressive performance on a wide variety of medical NLP task. However, there still lacks a large language model (LLM) specifically designed for cancer domain. Moreover, these LLMs typically have billions of parameters, making them computationally expensive for healthcare systems. Thus, in this study, we propose CancerLLM, a model with 7 billion parameters and a Mistral-style architecture, pre-trained on 2,676,642 clinical notes and 515,524 pathology reports covering 17 cancer types, followed by fine-tuning on three cancer-relevant tasks, including cancer phenotypes extraction, cancer diagnosis generation, and cancer treatment plan generation. Our evaluation demonstrated that CancerLLM achieves state-of-the-art results compared to other existing LLMs, with an average F1 score improvement of 8.1%. Additionally, CancerLLM outperforms other models on two proposed robustness testbeds. This illustrates that CancerLLM can be effectively applied to clinical AI systems, enhancing clinical research and healthcare delivery in the field of cancer. © 2024, CC BY.

关键词： Clinical research

Interpretable Differential Diagnosis with Dual-Inference Large Language Models

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zhou, Shuang Lin, Mingquan Ding, Sirui Wang, Jiashuo Melton, Genevieve B. Zou, James Zhang, Rui Division of Computational Health Sciences Department of Surgery University of Minnesota MinneapolisMN United States Bakar Computational Health Sciences Institute University of California San Francisco San FranciscoCA United States Department of Computing The Hong Kong Polytechnic University Hong Kong Hong Kong Institute for Health Informatics Division of Colon and Rectal Surgery Department of Surgery University of Minnesota MinneapolisMN United States Department of Biomedical Data Science Stanford University StanfordCA United States

Automatic differential diagnosis (DDx) is an essential medical task that generates a list of potential diseases as differentials based on patient symptom descriptions. In practice, interpreting these differential diagnoses yields significant value but remains under-explored. Given the powerful capabilities of large language models (LLMs), we investigated using LLMs for interpretable DDx. Specifically, we curated the first DDx dataset with expert-derived interpretation on 570 clinical notes. Besides, we proposed Dual-Inf, a novel framework that enabled LLMs to conduct bidirectional inference (i.e., from symptoms to diagnoses and vice versa) for DDx interpretation. Both human and automated evaluation validated its efficacy in predicting and elucidating differentials across four base LLMs. In addition, Dual-Inf could reduce interpretation errors and hold promise for rare disease explanations. To the best of our knowledge, it is the first work that customizes LLMs for DDx explanation and comprehensively evaluates their interpretation performance. Overall, our study bridges a critical gap in DDx interpretation and enhances clinical decision-making. Copyright © 2024, The Authors. All rights reserved.

关键词： Diseases

Integration of Droplet Microfluidic Tools for Single-cell Functional Metagenomics:An Engineering Head Start

学校读者我要写书评

暂无评论

Genomics, Proteomics & Bioinformatics 2021年第3期19卷 504-518页

作者： David Conchouso Amani Al-Ma’abadi Hayedeh Behzad Mohammed Alarawi Masahito Hosokawa Yohei Nishikawa Haruko Takeyama Katsuhiko Mineta Takashi Gojobori Department of Industrial and Mechanical Engineering Universidad de las Américas PueblaPuebla 72810Mexico Computational Bioscience Research Center King Abdullah University of Science and TechnologyThuwal 23955-6900Saudi Arabia Biological and Environmental Sciences and Engineering Division King Abdullah University of Science and TechnologyThuwal 23955-6900Saudi Arabia Research Organization for Nano&Life Innovation Waseda UniversityTokyo 162-0041Japan Department of Life Science and Medical Bioscience Waseda UniversityTokyo 162-8480Japan Institute for Advanced Research of Biosystem Dynamics Waseda Research Institute for Science and EngineeringWaseda UniversityTokyo 169-8555Japan Computational Bio Big-Data Open Innovation Laboratory AIST-Waseda UniversityTokyo 169–0072Japan Computer Electricaland Mathematical Sciences and Engineering DivisionKing Abdullah University of Science and TechnologyThuwal 23955-6900Saudi Arabia

Droplet microfluidic techniques have shown promising outcome to study single cells at high ***,their adoption in laboratories studying“-omics”sciences is still irrelevant due to the complex and multidisciplinary nature of the *** facilitate their use,here we provide engineering details and organized protocols for integrating three droplet-based microfluidic technologies into the metagenomic pipeline to enable functional screening of bioproducts at high ***,a device encapsulating single cells in droplets at a rate of~250 Hz is described considering droplet size and cell ***,we expand on previously reported fluorescence-activated droplet sorting systems to integrate the use of 4 independent fluorescence-exciting lasers(i.e.,405,488,561,and 637 nm)in a single platform to make it compatible with different fluorescence-emitting *** this sorter,both hardware and software are provided and optimized for effortlessly sorting droplets at 60 ***,a passive droplet merger is also integrated into our pipeline to enable adding new reagents to already-made droplets at a rate of 200 ***,we provide an optimized recipe for manufacturing these chips using silicon dry-etching *** of the overall integration and the technical details presented here,our approach allows biologists to quickly use microfluidic technologies and achieve both single-cell resolution and high-throughput capability(>50,000 cells/day)for mining and bioprospecting metagenomic data.

关键词： Droplet microfluidics Single cell Metagenomics Droplet sorter Biotechnology

Coupling streaming AI and HPC ensembles to achieve 100–1000× faster biomolecular simulations

学校读者我要写书评

暂无评论

Coupling streaming AI and HPC ensembles to achieve 100–1000...

International Symposium on Parallel and Distributed Processing (IPDPS)

作者： Alexander Brace Igor Yakushin Heng Ma Anda Trifan Todd Munson Ian Foster Arvind Ramanathan Hyungro Lee Matteo Turilli Shantenu Jha Data Science and Learning Division Argonne National Laboratory Lemont USA Department of Computer Science University of Chicago Chicago USA Data Science and Learning Division Argonne National LaboratoryLemontUSA Center For Biophysics University of Illinois Urbana-Champaign Champaign USA Mathematics and Computer Science Division Argonne National LaboratoryLemontUSA RADICAL-Lab Rutgers University New Brunswick USA Computational Science Initiative Brookhaven National Laboratory Upton USA

Machine learning (ML)-based steering can improve the performance of ensemble-based simulations by allowing for online selection of more scientifically meaningful computations. We present DeepDriveMD, a framework for ML-driven steering of scientific simulations that we have used to achieve orders-of-magnitude improvements in molecular dynamics (MD) performance via effective coupling of ML and HPC on large parallel computers. We discuss the design of DeepDriveMD and characterize its performance. We demonstrate that DeepDriveMD can achieve between 100-1000× acceleration for protein folding simulations relative to other methods, as measured by the amount of simulated time performed, while covering the same conformational landscape as quantified by the states sampled during a simulation. Experiments are performed on leadership-class platforms on up to 1020 nodes. The results establish DeepDriveMD as a high-performance framework for ML-driven HPC simulation scenarios, that supports diverse MD simulation and ML back-ends, and which enables new scientific insights by improving the length and time scales accessible with current computing capacity.

关键词： Proteins Couplings Computers Accelerometers Solid modeling Distributed processing computational modeling

Transferable Graph Neural Fingerprint Models for Quick Response to Future Bio-Threats

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Chen, Wei Ren, Yihui Kagawa, Ai Carbone, Matthew R. Chen, Samuel Yen-Chi Qu, Xiaohui Yoo, Shinjae Clyde, Austin Ramanathan, Arvind Stevens, Rick L. van Dam, Hubertus J.J. Lu, Deyu Center for Functional Nanomaterials Brookhaven National Laboratory UptonNY United States Computational Science Initiative Brookhaven National Laboratory UptonNY United States Data Science and Learning Division Argonne National Laboratory LemontIL United States Condensed Matter Physics & Materials Science Brookhaven National Laboratory UptonNY United States

Fast screening of drug molecules based on the ligand binding affinity is an important step in the drug discovery pipeline. Graph neural fingerprint is a promising method for developing molecular docking surrogates with high throughput and great fidelity. In this study, we built a COVID-19 drug docking dataset of about 300,000 drug candidates on 23 coronavirus protein targets. With this dataset, we trained graph neural fingerprint docking models for high-throughput virtual COVID-19 drug screening. The graph neural fingerprint models yield high prediction accuracy on docking scores with the mean squared error lower than 0.21 kcal/mol for most of the docking targets, showing significant improvement over conventional circular fingerprint methods. To make the neural fingerprints transferable for unknown targets, we also propose a transferable graph neural fingerprint method trained on multiple targets. With comparable accuracy to target-specific graph neural fingerprint models, the training and data efficiency of the transferable model is several times higher. We highlight that the impact of this study extends beyond COVID-19 dataset, as our approach for fast virtual ligand screening can be easily adapted and integrated into a general machine learning-accelerated pipeline to battle future bio-threats. © 2023, CC BY.

关键词： COVID-19

Structured barycentric forms for interpolation-based data-driven reduced modeling of second-order systems

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Gosea, Ion Victor Gugercin, Serkan Werner, Steffen W.R. Max Planck Institute for Dynamics of Complex Technical Systems Sandtorstr. 1 Magdeburg39106 Germany Department of Mathematics and Division of Computational Modeling and Data Analytics Academy of Data Science Virginia Tech BlacksburgVA24061 United States Courant Institute of Mathematical Sciences New York University New YorkNY10012 United States

An essential tool in data-driven modeling of dynamical systems from frequency response measurements is the barycentric form of the underlying rational transfer function. In this work, we propose structured barycentric forms for modeling dynamical systems with second-order time derivatives using their frequency domain input-output data. By imposing a set of interpolation conditions, the systems’ transfer functions are rewritten in different barycentric forms using different parametrizations. Loewner-like algorithms are developed for the explicit computation of second-order systems from data based on the developed barycentric forms. Numerical experiments show the performance of these new structured data driven modeling methods compared to other interpolation-based data-driven modeling techniques from the *** Codes 41A20, 65D05, 93B15, 93C05, 93C80 Copyright © 2023, The Authors. All rights reserved.

关键词： Interpolation

Structured vector fitting framework for mechanical systems

学校读者我要写书评

暂无评论

IFAC-PapersOnLine 2022年第20期55卷 163-168页

作者： Steffen W.R. Werner Ion Victor Gosea Serkan Gugercin Courant Institute of Mathematical Sciences New York University New York NY 10012 USA Max Planck Institute for Dynamics of Complex Technical Systems Sandtorstr. 1 39106 Magdeburg Germany Department of Mathematics and Division of Computational Modeling and Data Analytics Academy of Data Science Virginia Tech Blacksburg VA 24061 USA

We develop a structure-preserving formulation of the data-driven vector fitting algorithm for modally damped mechanical systems. Using the structured pole-residue form of the transfer function of modally damped second-order systems, we propose two potential, structured extensions of the barycentric formula for system transfer functions. Integrating these new forms within the classical vector fitting algorithm leads to the formulation of two new algorithms. These allow the computation of modally damped mechanical systems from data in a least-squares fashion, where the learned model is guaranteed to have the desired structure. We test the proposed algorithms on two benchmark models.

关键词： data-driven modeling mechanical systems reduced-order modeling vector fitting least-squares fit barycentric forms

Adapting Segment Anything Model for Ulcer Segmentation in Inflammatory Bowel Disease

学校读者我要写书评

暂无评论

Adapting Segment Anything Model for Ulcer Segmentation in In...

Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)

作者： Lingrui Cai Ryan W. Stidham Emily Wittrup Kayvan Najarian Department of Computational Medicine and Bioinformatics University of Michigan Ann Arbor Michigan Department of Internal Medicine Division of Gastroenterology Michigan Medicine Ann Arbor Michigan Michigan Institute for Data Science University of Michigan Ann Arbor Michigan Max Harry Weil Institute for Critical Care Research and Innovation University of Michigan Ann Arbor Michigan

ISBN: (数字)9798350371499

ISBN: (纸本)9798350371505

Inflammatory Bowel Disease (IBD) is a widespread gastrointestinal disorder characterized by inflammation and ulceration, affecting millions of individuals. Mucosal injury assessment through colonoscopy plays a crucial role in determining disease severity and therapeutic efficacy in IBD. This study draws inspiration from the Segment Anything Model (SAM) framework and tailors it for ulcer segmentation using a Parameter-Efficient Transfer Learning (PETL) approach and incorporating binary cross-entropy loss with boundary loss. Comparative assessments against state-of-the-art UNet-based convolutional neural network models underscore SAM’s superior performance, especially when faced with limited training samples. Our ulcer segmentation models demonstrate Dice, IoU, and 95%HD scores of 61.92±0.69, 75.18±0.38, and 10.07±0.05 in the colon dataset, and 55.60±0.82, 77.00±0.67, and 7.74±0.23 in the ileum dataset, respectively. This study highlights the potential of our approach to significantly enhance the efficiency and accuracy of IBD diagnosis and assessment through cutting-edge computer vision techniques.

关键词： Inflammatory bowel disease Training Adaptation models Accuracy computational modeling Transfer learning Colonoscopy Transformers Colon Convolutional neural networks

On the particle approximation of lagged Feynman–Kac formulae

学校读者我要写书评

暂无评论

Stochastic Processes and their Applications 2025年 188卷

作者： Awadelkarim, Elsiddig Caffarel, Michel Del Moral, Pierre Jasra, Ajay Applied Mathematics and Computational Science Program Computer Electrical and Mathematical Sciences and Engineering Division King Abdullah University of Science and Technology Thuwal23955-6900 Saudi Arabia University of Toulouse CNRS-Lab. de Chimie et Physique Quantiques Toulouse31062 France Centre de Recherche Inria Bordeaux Sud-Ouest Talence33405 France School of Data Science The Chinese University of Hong Kong Shenzhen Shenzhen China

In this paper we examine the numerical approximation of the limiting invariant measure associated with Feynman–Kac formulae. These are expressed in a discrete time formulation and are associated with a Markov chain and a potential function. The typical application considered here is the computation of eigenvalues associated with non-negative operators as found, for example, in physics or particle simulation of rare-events. We focus on a novel lagged approximation of this invariant measure, based upon the introduction of a ratio of time-averaged Feynman–Kac marginals associated with a positive operator iterated l∈N times;a lagged Feynman–Kac formula. This estimator and its approximation using Diffusion Monte Carlo (DMC) are commonly used in the physics literature. In short, DMC is an iterative algorithm involving N∈N particles or walkers simulated in parallel, that undergo sampling and resampling operations. In this work, it is shown that for the DMC approximation of the lagged Feynman–Kac formula, one has an almost sure characterization of the L1-error as the time parameter (iteration) goes to infinity and this is at most of O(exp{−κl}/N), for κ>0. In addition a non-asymptotic in time, and time uniform L1−bound is proved which is O(l/N). We also prove a novel central limit theorem to give a characterization of the exact asymptotic in time variance. This analysis demonstrates that the strategy used in physics, namely, to run DMC with N and l small and, for long time enough, is mathematically justified. Our results also suggest how one should choose N and l in practice. We emphasize that these results are not restricted to physical applications;they have broad relevance to the general problem of particle simulation of the Feynman–Kac formula, which is utilized in a great variety of scientific and engineering fields. © 2025 Elsevier B.V.

关键词： Iterative methods