notwa/smbot
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

fix non-antithetic case for ARS

Browse source
This commit is contained in:
Connor Olding 2018-06-13 22:46:09 +02:00
parent 601d78bfda
commit ac4c534185
1 changed files with 11 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

9
ars.lua
View file

@ -119,17 +119,22 @@ function Ars:tell(scored, unperturbed_score)
self.evals = self.evals + #scored self.evals = self.evals + #scored
if use_lips then self.evals = self.evals + 1 end if use_lips then self.evals = self.evals + 1 end
-- FIXME: the non-antithetic case seems to be broken.
local indices = collect_best_indices(scored, self.poptop, self.antithetic) local indices = collect_best_indices(scored, self.poptop, self.antithetic)
local top_rewards = {} local top_rewards = {}
for i = 1, #scored do top_rewards[i] = 0 end for i = 1, #scored do top_rewards[i] = 0 end
if self.antithetic then
for _, ind in ipairs(indices) do for _, ind in ipairs(indices) do
local sind = (ind - 1) * 2 + 1 local sind = (ind - 1) * 2 + 1
top_rewards[sind + 0] = scored[sind + 0] top_rewards[sind + 0] = scored[sind + 0]
top_rewards[sind + 1] = scored[sind + 1] top_rewards[sind + 1] = scored[sind + 1]
end end
else
for _, i in ipairs(indices) do top_rewards[i] = scored[i] end
-- note: although this normalizes the scale, it's later
-- re-normalized differently by reward_dev anyway.
top_rewards = util.normalize_sums(top_rewards)
end
local step = nn.zeros(self.dims) local step = nn.zeros(self.dims)
local _, reward_dev = calc_mean_dev(top_rewards) local _, reward_dev = calc_mean_dev(top_rewards)