News for hackers. Stuff that matters.
Re:Machine learning is fundamentally conservative by TeMPOraL on Saturday January 04, 2020 @02:14AM (#21953436)
Perhaps you could call it "integral controller" (like the I in "PID controller")? Because systems that have a memory behave like ones, and we are definitely in a feedback loop with those systems.

And, from what I remember from my control theory classes, the integral part of a controller introduces lag, inertia, generally making the output more resistant to input changes.

(Also note that the "non-conservative" Proportional and Derivative components in a PID by definition don't learn - they react to input and its change.)

Re:Machine learning is fundamentally conservative by reilly3000 on Saturday January 04, 2020 @02:31AM (#21953504)
That's a pretty interesting corollary! I suppose I could extrapolate that in such a system, too much change over too little time would invite a strong conservative response / lag to future input.

Our 20 year old millennium has seen a tremendous rate of change in both technology and social norms. Perhaps politics has some similar inertial dynamics.