Is this a task an AI can learn?

We are your partner for AI testing.

Testing AI with a proven method

Tap into our experience from a wide range of successful AI testing projects in various industries such as automotive, construction, medical, waste management, and sewing. Our blueprint for Computer Vision testing aims at providing a reliable proof of concept at minimum investment costs – suitable for use cases from the circular economy but also new part production.

W. Klat & H. Brandt-Pook (2021): “A Blueprint for Computer Vision Testing in the Circular Economy”

In: V. Wohlgemuth et al.: “Environmental Informatics. A bogeyman or saviour to achieve the UN Sustainable Development Goals?” Proceedings of the 35th edition for the EnviroInfo: 251-258.

Our two hypothesis AI testing method

Hypothesis 1

The task is solvable based on solely digital image data.

Hypothesis 1 evaluates whether visual image data is sufficient for task completion and if the image data can be acquired in the production process. We capture sample data and test digital task completion with your employees.


Duration: 2-3 weeks

Hypothesis 2

An open source deep learning algorithm is able to learn the task.

Hypothesis 2 tests the compatibility of state-of-the-art neural networks with your task. We create a first training dataset and evaluate the training progress of neural networks.

 

Duration: 3-6 weeks

Are you ready for your first AI test?

Let's discuss your use case!