Advances in machine learning – particularly deep learning – are fuelling the belief that AI has come of age.
But there are sizeable barriers to widespread deployment. Firstly, the requirement for huge, high-quality data sets to train deep learning systems. Secondly, the resulting 'AI black box', making it difficult to establish trust.
This talk will detail creative, technology-based approaches we can take to overcome these barriers. For example, approaches that require far less source data, or where we can ask an AI system to explain its decisions. We will illustrate the talk with results from our Digital Greenhouse AI research laboratory.