Rao and Ballard, 1999
Karl Friston thinks this is the modern paper that “started it all.”
Karl Friston, 2010
This 2010 Karl Friston paper, while largely incomprehensible to the mathematically naive, played a major role in putting PC “on the map”
This 2017 tutorial by Bogacz is very helpful for working through the Friston 2010 mathematics (above). It is more complicated than a high-school education allows, but is clearly easier to understand than Friston’s maths. Budgeting around 8-12 hours to work through it will yield excellent results.
Here Williams, a philosopher, raises serious doubts about whether predictive coding is capable of doing all that it claims to do.