1 of 7

making sense of text and data

Sirma AI (Ontotext)

September 2021

https://www.ontotext.com/

2 of 7

Ontotext

Products

Sectors

Projects

Sirma AI (trading as Ontotext) is established in 2000, as a Research Lab of Sirma Holding.
It provides unique mix of knowledge graphs and advanced NLP techniques.
Ontotext is a leader in semantic web technology and AI
The main products of the company are

GraphDB – one of the leading RDF triple stores, providing full stack of tools and services for knowledge graph managements
The Ontotext Platform vision, technology, and business are about making sense of text and data. Its ambitions are to let big knowledge graphs improve the accuracy of text analytics, to use text analytics to interlink and enrich knowledge graphs, and to allow knowledge graphs to enable better search, exploration, classification, and recommendation across diverse information spaces.

The company has a solid experience in research projects - more than 50 research and innovation projects

some of the current H2020 projects and the respected sectors are listed here: Life science and healthcare; Food & agriculture; financial services; media and publishing; transport and logistics; industry, including manufacturing, automotive, energy, bulging automation, infrastructure and aerospace.

Ontotext Expertise

Ontotext Expertise

Project idea: Semantization of DS & ML

We've had this project idea for a year now, following some initial engagements with a large car maker and the US Air Force:

Problem: Data Science, AI & ML are expensive, and that's one of the reasons why relatively few enterprises use it.

Goal: rationalize and industrialize DS efforts, and make them more reproducible and reusable.

Approach: capture a lot of semantic info about all DS processes in an enterprise, and thus enable automation, discovery, reusabulity.

Left box: kinds of data we'd like to represent and integrate semantically; part of it is similar to what you can see on the Kaggle and OpenML sites:

Business context: goals, motivations, data value, value chain, cost vs benefit analysis, SWOT analysis...
DS challenges, where do they come from, datasets that can be leveraged to solve them
DS staff, expertise, projects, tasks, risks
DS/ML algorithms, implementations, modules, dependencies, software projects, versions, issue trackers
Cloud and IT resources: compute, storage; their deployment, management, automation...
ML model deployment, performance, model drift, retraining…

Right box: Established software genres that cover parts of this landscape:

ModelOps (devOps for ML), Feature Spaces
Enterprise data catalogs (data hubs) vs data marketplaces vs open data catalogs vs EU Data Spaces and their metadata
FAIR data, reproducible research, Research Objects, research workflows

Potential Data Scope