https://store-images.s-microsoft.com/image/apps.38799.7d27102e-dd63-4129-b018-bc9040cb0955.09c697dd-7377-441b-a844-2da55ca3cbf4.4a00bb1a-e6f2-4281-bee0-fdfc043f79c2

John Snow Labs - Healthcare NLP

John Snow Labs Inc

(1 ratings)

John Snow Labs - Healthcare NLP

John Snow Labs Inc

(1 ratings)

NLP & OCR libraries, models and notebooks for text and image annotation and model training & tuning

Overview

    John Snow Labs’ exclusive Healthcare NLP and OCR Libraries are curated to extract meaningful insights from unstructured documents, amplifying the efficiency and accuracy of data interpretation in healthcare.

About the offer

    This package includes specialized Natural Language Processing (NLP) Python libraries, explicitly curated for the Healthcare sector. It simplifies and accelerates text annotation processes, leveraging pre-trained Deep Learning (DL) models, rules, and effective prompts, in addition to facilitating NLP model evaluation, training, and enhancement. The product assures unparalleled accuracy, intrinsic scalability, advanced hardware optimizations, and effortless compose text and image processing pipelines.

    This subscription does not impose any restrictions on the number of documents, models, or pipelines utilized. Licensing is administered on a per-server basis.

    For enterprise licenses contact us here

What is included

    Spark NLP is a popular enterprise-level NLP library. It offers robust, scalable, and updatable features based on the newest NLP research, along with over 10,000 pre-trained models for various tasks like Entity Recognition, Information Extraction, and Emotion Detection, among others.

    The Healthcare NLP tools and models specialize in clinical and biomedical Named Entity Recognition, identifying over 400 entity types and determining assertion statuses, which helps in extracting and correlating clinical information, referencing it to standard codes like SNOMED-CT and ICD-10.

    The Visual NLP (OCR) tools are equipped to understand and process forms and images, focusing on tasks such as table detection, image enhancement, and visual document classification, ensuring comprehensive text and image analysis and de-identification.

    Gain full access to an extensive range of models and pipelines available on the NLP Models Hub, including over 1,200 specialized in healthcare and 17,500 general models. Benefit from 30+ ready-to-use Jupyter notebooks, facilitating an easy start with essential NLP tasks such as text classification, sentiment analysis, and named entity recognition, relation extraction, assertion status, entity linking, de-identification, translation, summarization, question answering, spelling and grammar.

    This package also comes with a user-friendly NLU python library, simplifying text understanding and model testing with a straightforward coding approach. Additionally, benefit from the Spark NLP Display library, which enriches your text with intuitive and helpful annotations. This offer is crafted for those who seek comprehensive yet accessible tools for text and image analysis in NLP.

This offer is tailored for:

  • Python developers aiming to work with text, images, and PDF documents for entity and relation extraction.
  • Data scientists addressing challenges in natural language processing (NLP).
  • Machine learning engineers involved in the testing, training, and tuning of NLP models.
  • Research groups focusing on deriving insights from unstructured natural language documents.
  • Individuals or groups interested in text and image analysis, data extraction, document labeling, or NLP model training.

Target verticals

    The Spark NLP and Visual NLP libraries in this package are versatile, working across various domains and supporting documents in over 250 languages. This offer includes a specialized Healthcare library, with its pre-trained models exclusive to healthcare. The NLP Models Hub is comprehensive, featuring over 17,000 pre-trained models for a variety of purposes and an additional 1,200+ models specifically optimized for healthcare-related tasks.

Technical Specifications

  • Recommended memory: 32GB RAM
  • Recommended vCPU:8 vCPUs
  • Operating System:Ubuntu 20.04

Included integrations

  • Jupyter Lab preinstalled and available at http://{IP_ADDRESS}/jupyter. The password to login to Jupyter Lab is the SUBSCRIPTION_ID

3 Easy Steps to get started

  1. Subscribe to the product on the Azure Marketplace.
  2. Deploy it on a new machine.
  3. Access the welcome page for a guided experience on http://instanceIP
https://store-images.s-microsoft.com/image/apps.57165.7d27102e-dd63-4129-b018-bc9040cb0955.9cde186a-29ce-4d00-bd00-1d9825150e0d.6785f24b-92d4-446c-87c0-2924dccf71cd
/staticstorage/8165fe0/assets/videoOverlay_7299e00c2e43a32cf9fa.png
https://store-images.s-microsoft.com/image/apps.57165.7d27102e-dd63-4129-b018-bc9040cb0955.9cde186a-29ce-4d00-bd00-1d9825150e0d.6785f24b-92d4-446c-87c0-2924dccf71cd
/staticstorage/8165fe0/assets/videoOverlay_7299e00c2e43a32cf9fa.png
https://store-images.s-microsoft.com/image/apps.59931.7d27102e-dd63-4129-b018-bc9040cb0955.09c697dd-7377-441b-a844-2da55ca3cbf4.be403b4b-78b4-4d3a-b51e-5dbc338183d3
/staticstorage/8165fe0/assets/videoOverlay_7299e00c2e43a32cf9fa.png
https://store-images.s-microsoft.com/image/apps.29373.7d27102e-dd63-4129-b018-bc9040cb0955.21d2558c-0f58-472e-9bd3-0d1093a68034.c9d78942-c96a-4843-b98c-121a5a671f8c
/staticstorage/8165fe0/assets/videoOverlay_7299e00c2e43a32cf9fa.png
https://store-images.s-microsoft.com/image/apps.36724.7d27102e-dd63-4129-b018-bc9040cb0955.09c697dd-7377-441b-a844-2da55ca3cbf4.2b84b54b-4d38-4fe1-8b56-d0e6642959af
https://store-images.s-microsoft.com/image/apps.31852.7d27102e-dd63-4129-b018-bc9040cb0955.09c697dd-7377-441b-a844-2da55ca3cbf4.4e993f4d-92a8-4517-bd57-a4560f4227a1