Project semANT
We are developing new ways how to navigate, search, explore, visualize and analyze large collections of text documents. We are defining how digitized printed and handwritten documents will be used and accessed in the future.
Make sense of documents
Czech libraries and archives are home to a wealth of digitized documents, and we’ve seen significant strides in improving their online accessibility and search functions. However, the current full-text search, while adept at recognizing different word forms, lacks the ability to understand their meaning, making it a challenge to find documents on specific topics.
At semANT, we’re dedicated to simplifying the search process and improving the exploration of digitized documents, making it more intuitive and insightful for everyone.
We are working with
Our Goal
What are we working on
We are creating a next generation document access application incorporating advanced AI capabilities.
Semantic search
We are building semantic text embedding models for historic documents which allow us to retreive thematicaly relevant texts.
Adaptive search
Search results should adapt to current user goals and recent history. The search system should propose meaingfull query refinements and filters.
Question answering
Natural language interaction is the future. We strive to combine document retrieval with chatbot-like interface to efficiently anser user questions.
Topic discovery
Discover topics in collection of documents or search results. Tag user topics, analyze and shere them.
Visualizations
Visualize topics developing in time, topic contexts, topic relations and content.
Languages
Seamlessly search across documents from many languages, and explore the documents in your language using state-of-the-art automatic translation.