Tag Archives: McNeill

Incorporating prior knowledge for inference (PhD)

One of the key challenges in pattern classification, clustering and regression is the ability to make maximal use of prior knowledge. Popular machine learning (ML) techniques are often ineffective when applied to real world problems, in many cases, due to the naive use of unsuitable distance measures that are not necessarily justified based on prior information. This project will look at principled ways of incorporating prior knowledge through development of task dependent distance metrics.

Related Themes

Related People