Knowledge Integration

CIIM AI Framework launched

In collaboration with the Horniman Museum and Gardens and the Science Museum Group, we have developed an AI framework for our CIIM product. It works with numerous third party AI services, such as Google Vision, and also provides integration with the GATE (General Architecture in Text Engineering) toolkit from Sheffield University.

Image by Gerd Altmann

Artificial Intelligence services such as Google Vision and Amazon Comprehend are being used to enrich data to bring about a diverse range of benefits. Integrating these services with CIIM provides exciting possibilities. Our customers will be able to take advantage of the many of benefits of AI, to support delivery of:

  • Enhanced description of images and artifacts for superior search
  • Creation of new connections between artifacts
  • Identification of patterns in data
  • Natural language processing services
  • Automation of tasks to save staff time

Multiple AI services can now be added to an organisation’s CIIM platform. They can be configured to work independently on different aspects of the metadata or to work together: either in a pipeline with the output of one service feeding into another, or collaboratively with data passed in and out of services in a cycle of enrichment.

Resulting enhancements can automatically be added to the core data or quarantined first to allow them to go through a review workflow.

Many AI service providers charge for use of their APIs but tend to have a “free tier” offering a maximum number of transactions over a fixed period. To help manage costs the CIIM AI framework can be configured to monitor use and ensure subscriptions stay within the free tier of use.

Horniman Museum and Gardens uses Dokuwiki pages to provide narrative content to their online collections. Our integration of the GATE toolkit enables analysis of Dokuwiki page content and automatic alignment of the pages with the internal thesauri (derived from CIIM via extraction from their collections management system – MIMSY XG).

At the Science Museum the framework is being used to provide integration with Amazon Rekognition to analyse image content and add this to the object and archive records online.

CIIM currently integrates with the below AI services, and we expect this list to grow as new services become available:

  • Google Vision (image analysis)
    • Labels
    • Faces
    • Text in Images
    • Web Entities
    • Landmarks
    • Localisation
    • Image Properties (e.g dominant colours)
  • Google Place (text analysis)
    • Geo-location
    • Organisation Details/Opening Hours
  • Amazon Rekognition (image analysis)
    • Faces
    • Labels
    • Text in images
  • Amazon Comprehend (text analysis)
    • Entities
    • Key Phrases
    • Sentiment
    • Syntax
  • IBM Watson Natural Language Understanding (text analysis)
    • Entities
    • Keywords
    • Concepts
    • Categories
  • GATE (text analysis)
    • Supply training data
    • Supply thesaurus data
    • Thesaurus alignment
    • Custom annotation parsing
  • Clarifai (image analysis)
    • Apparel
    • Image properties (e.g. dominant colours)
    • Face detection
    • Textures and Patterns

The AI framework is now part of the core CIIM product and is available to all users. Get in touch if you would like to know more.