aidkit is the testing solution for your data collection and safety validation  

The road out there is filled with challenges and hazards, all which your autonomous vehicle needs to navigate safely. But making your ML perception system safe can entail an expensive cycle of data collection and testing.

Data augmentation with aidkit breaks the cycle and gets your processes flowing through comprehensive testing and guidance for lean data collection and labelling that saves you time and money.

No matter the model, scene, or ODD, aidkit ’til you make it.

Harmonize cutting-edge AI with the safety we've come to demand of the automobile

aidkit is our software solution for testing the perception functions of ADAS/ADS in order to ensure they are safe for deployment.

By scalably testing your ML model for robustness against a range of augmented scenes, aidkit ensures reliable performance in your ODD.

Where performance falters, aidkit guides your data campaign so you know how to improve and don't waste time and money on unfruitful collection.

Deploy a safer perception component with the empirical evidence to back up your safety claims to managers, regulators, and customers.

Scalable perception component testing

aidkit generates a wide range of images augmented with the conditions you specify. The scalable testing pipeline runs within the MLOps workflow you choose with the parameters you set. Quickly get the performance- and worst-case analysis you need.

Lean and targeted data campaigns

Testing with aidkit allows you to easily identify where your model needs more data for retraining. Its ODD-centric design ensures you can conduct tailored data campaigns that meet your needs. Fill your gaps without expensive, extraneous collection and labeling.

Explainable and evidenced safety arguments

aidkit provides you with evidence-based risk scores showing probability and severity for every tested ODD scenario. Results can be seen in our reporting feature, which can feed into your EU AI Act compliance.  Prove your perception model safely handles all the required cases in your ODD.

aidkit: trust in your AI begins with testing you can trust

By augmenting data with perturbation-based image operations, aidkit tests the performance of perception models under the various conditions possible in an ODD but not easily captured in fleet data. Along with real-world data for training and simulation for later testing, augmented data should be a pillar in your methodological triangulation: the key to a strong scientific basis and confidence in your AI models.

aidkit makes these complex ideas and tasks easy to implement and understand. It integrates into existing workflows and reduces time to deployment though useful features such as scaling, data lineage, and automatic retraining. With aidkit you can translate the requirements of current and upcoming safety standards into technical pass/fail criteria and then clearly communicate you've met the criteria.

Driving safe perception

Check out our safety use case to learn about how testing with aidkit work in a real-world situation.

© neurocat GmbH