Optimizing data profiling in artificial intelligence projects

Discover

As part of the Confiance.ai program, IRT SystemX has developed DebiAI, an open-source data profiling tool designed to enhance dataset quality and improve the performance of artificial intelligence (AI) models.

The quality of datasets and the performance of models are critical to the success of R&D and industrial projects that rely on AI technologies. DebiAI addresses the growing needs of industrial stakeholders for effective use and control of AI models by offering a comprehensive solution for data analysis and instant bias detection. The tool enables data scientists to improve the accuracy and efficiency of their models, facilitating more informed decision-making based on trusted datasets.

DebiAI has established itself as a benchmark tool for exploring complex datasets and evaluating models in context. Tested on several use cases—including those led by Thales and Safran—DebiAI has proven its effectiveness in demanding industrial environments. Its strength lies in its open-source nature, which makes it easy to adopt and integrate. A public version of the tool is available for demonstration and hands-on use.

DebiAI is more than just a data analysis tool. It’s a performance catalyst for AI projects, part of the new family of HCAI (Human-Centered Artificial Intelligence) tools. It helps explore complex datasets and optimize models for reliable and robust results, all in a user-friendly way.

Raphaël Braud, Head for the Data Science and AI team, IRT SystemX

Digital security and networks