hmc-cs-mdrissi / neural_nets_research

A collection of projects for neural nets research in Mehdi's research group.
MIT License
7 stars 1 forks source link

Using the Halting Probability in the ANC for all instructions #14

Open hmc-cs-mdrissi opened 6 years ago

hmc-cs-mdrissi commented 6 years ago

Currently if the anc is in a half half state it'll continue to do all other instructions normally. It may be worth instead having the anc do all other instructions with only half probability. This is equivalent to adding one additional line of code to the program with instruction halt with probability 1 and treating halt jump to that line.