About CAKT

Printer-friendly versionSend by email

Content Analysis and Knowledge Technologies Group


In the era of overinformation and content abundance, research faces the challenge of providing actionable insights and knowledge from the underlying wealth of information. The challenge is directly related to whether we can efficiently query, analyse, interprete and visualize data in a way that will be usable by both experts and non-experts. The multi-modal nature of the data people share and interact upon poses another difficulty to the analysis process. Essentially, in order to face the challenge, experts from different disciplines need to join forces.

Given this real need, the CAKT group brings together researchers and developers covering a variety of domains related to Natural Language Processing, Data Mining, Machine Learning, Knowledge Representation and Visualization under a common aim. The aim is to develop top-notch, scalable and widely adoptable methods and system solutions for multi-modal content analysis (e.g. text, image, speech, video), knowledge representation and management.

Below we provide a one-glance overview of the group.

CAKT overview

In the following paragraphs we elaborate on the vision and goals of the CAKT group.


CAKT aims to:

  • Achieve robust, efficient and accurate analysis for text and data.
  • Break the language barriers, especially for under-resources languages, by applying language-neutral analysis methods were needed, while harvesting existing resources where available.
  • Allow responsive, large-scale application of text and data analysis in real settings.


In CAKT, we strive to

  • Create tools to annotate and enrich existing content automatically, in order to add value and increase the usefulness of the .
  • Provide ways to combine and exploit varying data sources, by discovering, aligning and enriching available data, whether they be in structured, semi-structured or  unstructured form.
  • Implement tools that can render the analysis results as friendly and expressive data summaries, aiming to support data-related challenges faced by non-IT people in a variety of practices (e.g. policy making, reputation management, e-participation, data journalism).
  • Bring state-of-art research to the society, empowering e-participation, policy making, data journalism, reputation management.

Content analysis methods exploitable in various, multi-modal settings
Scalable data processing to respond to current analysis needs
Open tools and resources empowering the industry and the society


In CAKT we combine expertise and know how on a variety of domains, related to:

  • Knowledge extraction, management and use
    • Ontology learning, evolution and mapping
    • Information extraction
    • Metadata and knowledge description
    • Knowledge engineering
    • Open data, Linked data
    • Distributed, scalable querying over knowledge bases
  • Natural language processing and text mining
    • Adding domain and social knowledge to content
    • Automatic Summarization and Summary Evaluation
  • Pattern recognition, Data mining, Machine learning
  • Bioinformatics

© 2019 - Institute of Informatics and Telecommunications | National Centre for Scientific Research "Demokritos"

Terms of Service and Privacy Policy