take a number of observations, as he suggests, and let the machine learn the code by determining the coefficients A₀ etc. by simple curvilinear regression. The duration of previous and subsequent signals and the interpretation given to prev…I maintain that there is some merit in studying self-improvement systems and if I am going to do that I am going to study systems where I know what I want, rather than more difficult problems, however attractive they may be to mathematician…
Donald MacKay
scientist · 3 mentions across 2 readings
In this course
Donald MacKay was a pioneering cyberneticist and neuroscientist who developed probabilistic and information-theoretic approaches to understanding perception and self-regulating systems. The course readings draw on his work to ground discussions of how machines might learn through statistical inference and adaptive feedback, positioning his ideas about goal-directed self-improvement systems as foundational to cybernetic thinking about artificial intelligence. His insistence on studying well-defined optimization problems rather than abstract mathematical puzzles establishes a practical epistemology for the machine learning methods discussed in the seminar.
Mentioned in 2 readings
Appears alongside
People mentioned in the same passages — sorted by co-occurrence weight.