Embracing Uncertainty: The New Machine Intelligence

Many of the early applications of machine intelligence were based on expert systems constructed using rules elicited from human experts. Limitations in the applicability of this approach helped to drive black-box statistical methods, such as neural networks, based on learning from data. These black-box methods too are hitting limitations, due to the challenges of incorporating background knowledge. In this talk I will describe a new paradigm for machine intelligence which has emerged over the last five years, and which allows prior knowledge from domain experts to be integrated with machine learning techniques to enable a new generation of large-scale applications. The talk will be illustrated with tutorial examples as well as real-world case studies.