CLOSE

Atypon R&D Yields 4 Awards in Biomedical Semantic Indexing and Question Answering Competition

Our favorite
two days of the year:

Atypon Engage User Conferences

February 16, 2018

For the fifth year in a row, Atypon has placed in the International BioASQ Awards competition, this time winning awards in biomedical indexing, semantic question answering (QA), and funding information extraction.

Atypon’s R&D in machine learning and natural language processing—two branches of artificial intelligence—contributed to the award-winning semantic technologies, some of which underlie current and planned Literatum functionalities:

  • Semantic indexing: The technology behind Literatum’s auto-tagger.
  • Semantic search: The mechanism that will enable users to ask natural-language questions rather than search by keyword.
  • Funding extraction: A means of enriching article metadata with information about research funding.

BioASQ’s goal is to further the development of technology that solves information access problems in biomedical research. It organizes challenges for technology companies in biomedical semantic indexing, hierarchical text classification, machine learning, information retrieval, QA from texts and structured data, and multi-document summarization.

The Atypon team members participating in the challenge were Hong Zhou, Senior Product Manager, R&D; and Yannis Papanikolaou, Research Scientist.

They were joined by professors Grigorios Tsoumakas and Ioannis Vlahavas, and doctoral candidates Eirini Papagiannopoulos and Dimitris Dimitriadis, all of Aristotle University in Thessaloniki. Atypon funds several ongoing R&D projects at the university that explore ways to apply artificial intelligence to Literatum functionalities.

Atypon’s wins by category are:

  • Biomedical Semantic QA category, “concepts” task: 3 second-place finishes
  • Biomedical Semantic QA, “exact answers” task: 1 second-place finish
  • Large-scale Online Biomedical Indexing, flat-hierarchical measures: 3 second-place finishes
  • Funding Information Extraction from Biomedical Literature, grant-related task: 1 second-place finish

Take the next step.

Talk to us about migrating your publications to Literatum.