Craig Innes
Postdoctoral Research Associate
Contact:
- craiginnes [at symbol] gmail [dot symbol] com
- Website
Research Interests:
Expert-Assisted Decision Making, Unawareness, Symbolic-Probabilistic Hybrid Systems, Explainable AI, Graphical Models and Program Induction
Research Description:
As part of the sAIfer-surgery project, I am interested in learning richly structured, interpretable models and domain level specifications (i.e., a combination of logic/programs and dynamical systems) from data. In particular, the aim is to see if this can be done interactively. For example, can an automated assistant identify that a surgeon is about to proceed into the next phase of an operation? Can it offer the appropriate tool in response? Further, could such an assistant assess the technical skill of a practitioner or anticipate when they are about to make a mistake? How should such an assistant react if their actions are corrected?