CLAMP Core Products


CLAMP CMD

The CLAMP CMD(Command Line) is an off-the-shelf general-purpose clinical NLP system built on proven methods with good performance and high speed for extracting general clinical concepts such as diseases, drugs, labs, and procedures.

  • CLAMP components are built on proven methods in many clinical NLP challenges
  • Identify context of entities such as negation, certainty, and values
  • Support Multi-threads/Multi-processes/Docker/Distributed systems
  • Process local files/Database tables/HDFS files

CLAMP GUI

The CLAMP GUI(Graphical User Interface) system provides not only high-performance NLP modules and pipelines customized for clinical text, but also user-friendly interfaces for quickly building customized NLP pipelines. It provides multiple useful features for NLP professionals.

  • Annotate local data and re-train machine learning models with a few clicks
  • Specify rules on the top of ML models to improve performance
  • Export NLP pipelines as web services for easy integration
  • Provide Java APIs for individual components for system integration

CLAMP Enterprise

The CLAMP Enterprise is an server-based multiuser enterprise solution for NLP needs in healthcare organizations. User-friendly task scheduling, resource managing, indexing/searching, and visualization functions are supported in this version.

  • Schedule processing tasks
  • Allocate computational resources for each task
  • Search and visualize outputs of NLP outputs
  • Support Common Data Models (CDM) such as OMOP CDM

DeepMed-Deep Learning for Clinical NLP


DeepMed Docker

Our deep learning solutions - DeepMed provides a docker container that allows for an easy-to-use training of deep-learning models with a library of state-of-the-art algorithms(e.g., the BERT models) and automated parameter optimization.

DeepMed for SaaS

DeepMed is also available as a Software as a Service(SaaS) platform for Cloud-based computing with integrated state-of-the-art deep learning algorithms and improved Interoperability. It greatly shortens the time and reduces cost for clinical NLP development.

Melax Marketplace


Melax Knowledge Sharing Marketplace

The Melax Knowledge Sharing Marketplace allows users to share pipelines developed by themselves or to purchase pipelines of their interest. These pipelines include Cancer pipeline, Substance Overdose pipeline, Social Determinants pipeline, and many more...

Why CLAMP


High-performance NLP components built for clinical text:

  • Built on proven methods and models trained using clinical corpora.
  • Winning algorithms from many clinical NLP challenges such as i2b2 clinical NER (2009/2010-#2), SHARE/CLEF (2013-#1), and SemEval2014 UMLS encoding (#1).

Machine learning and hybrid approaches

  • Train your own models using machine learning and deep learning algorithms with a simple click
  • Compare different models with provided evaluation programs and visuallzation interfaces

Annotation and corpora management

  • Built-in annotation tool that supports different IO methods (e.g. files or databases) and collaborative annotation by a team
  • Compatible/convertible between different annotation tools (e.g., Brat, eHost, MAE etc.)

Customized pipelines

  • Build your own NLP pipelines using both machine learning and rules
  • Leverage existing pipelines to avoid NLP development from scratch

Knowledge bases and sample clinical text:

  • Provide all the knowledge bases (e.g., dictionaries) required for CLAMP components
  • Come with over 300 simulated clinical notes annotated with entities and relations

Interoperability and Scalability

  • Build on the UIMA framework, thus compatible with other systems such as cTAKES
  • Thread safe and validated on parallel computing platforms such as SPARK and Hadoop