Machine-Learning Based Clinical Decision Support

Human and Machine Learning

I wrote a book chapter for a Springer publication on the importance of trust and transparency in designing machine-learning systems for clinical decision support. Transparency is a major issue for some approaches such as neural nets, but even for those algorithms whose outputs are amenable to inspection, designing the interface to communicate the provenance of their decisions and suggestions is a major challenge. There are a number of human cognitive biases to consider in designing such interfaces, such as the phenomena of automation bias or risk homeostasis.

Read more in the full chapter, included here: Gretton Trust and Transparency in ML-CDSS

Human and Machine Learning

Human and Machine Learning

 

LEAVE A COMMENT