Testing Considerations for AI and ML in Wireless Networks
Artificial Intelligence (AI) and Machine Learning (ML) are core technologies that will pervasive and expansive in 6G. Thinking about how to test AI/ML communications systems present new challenges compared to traditional methods. There are 3 different types of AI/ML testing to consider—datasets (recorded from real world scenarios or artificially generated) are needed to train the AI/ML, AI/ML based test can be applied to wireless networks (for example, an AI/ML based signal identification), and finally testing to look at the performance of the AI/ML algorithm itself at a subcomponent level and at a system level. This session will dive into each of these three areas with a goal to leave attendees with a better understanding what test and validation means for AI and ML in wireless systems.