DELTA

— for safer AI decisions in industry

Introduction

The DELTA research project aims to strengthen trust in AI-based decision support in industry through increased reliability and transparency.

Purpose

A prerequisite for the willingness to use AI in decision-making is of course that the recommendations the system provides are correct. This requires not only a high level of accuracy but also that the AI system can assess its own uncertainty, how confident it is in its forecasts and suggestions. In addition, for the user to trust the AI, it needs to be able to explain how its decisions are made.

Expected results

The DELTA project is developing new models that can more accurately assess their own uncertainty and are easier to interpret through a transparent design. The work to ensure the robustness of AI-based decision-making is carried out using advanced methods such as conformal prediction and in collaboration with several industrial companies to adapt the solutions to real-world needs.

The project is expected to deliver innovative AI methods that improve decision-making in industry with, among other things, new models and algorithms to quantify uncertainty, increase explainability, and make AI easier to interpret in a tailored way.

Technical advances will be combined with user-friendly methods and concrete prototypes for easier evaluation and assessment of AI support.

By improving the credibility of AI, DELTA will contribute to more informed, trustworthy and transparent AI-supported decisions in industry. In Småland, Sweden and in the world.

DELTA is an acronym for Designing, Evaluating, and Leveraging Trustworthy AI solutions

Want to know more about the project?

Are you interested in a future collaboration?