def alice(preference, depth):
"""
Alice decides where to go by reasoning about Bob's choice
"""
alice_prior = location(preference)
with poutine.block():
bob_marginal = HashingMarginal(Search(bob).run(preference, depth - 1))
return pyro.sample("bob_choice", bob_marginal, obs=alice_prior)
def alice_fb(preference, depth):
"""
Alice's actual decision process:
Alice decides where to go by reasoning about Bob's choice
and choosing the other location.
"""
alice_prior = location(preference)
with poutine.block():
bob_marginal = HashingMarginal(Search(bob).run(preference, depth - 1))
pyro.sample("bob_choice", bob_marginal, obs=alice_prior)
return 1 - alice_prior
def bob(preference, depth):
"""
Bob decides where to go by reasoning about Alice's choice
"""
bob_prior = location(preference)
if depth > 0:
with poutine.block():
alice_marginal = HashingMarginal(Search(alice).run(preference, depth))
return pyro.sample("alice_choice", alice_marginal, obs=bob_prior)
else:
return bob_prior
def main(args):
# Here Alice and Bob slightly prefer one location over the other a priori
shared_preference = torch.tensor([args.preference])
alice_depth = args.depth
num_samples = args.num_samples
# We sample Alice's true choice of location
# by marginalizing over her decision process
alice_decision = HashingMarginal(Search(alice_fb).run(shared_preference, alice_depth))
# draw num_samples samples from Alice's decision process
# and use those to estimate the marginal probability
# that Alice chooses their preferred location
alice_prob = sum([alice_decision()
for i in range(num_samples)]) / float(num_samples)
print("Empirical frequency of Alice choosing their favored location " +
"given preference {} and recursion depth {}: {}"
.format(shared_preference, alice_depth, alice_prob))
def main(args):
# Here Alice and Bob slightly prefer one location over the other a priori
shared_preference = torch.tensor([args.preference])
bob_depth = args.depth
num_samples = args.num_samples
# We sample Bob's choice of location by marginalizing
# over his decision process.
bob_decision = HashingMarginal(Search(bob).run(shared_preference, bob_depth))
bob_prob = bob_decision._dist_and_values()[0].probs
print("bob prob", bob_prob)
# draw num_samples samples from Bob's decision process
# and use those to estimate the marginal probability
# that Bob chooses their preferred location
bob_prob = sum([bob_decision()
for i in range(num_samples)]) / float(num_samples)
print("Empirical frequency of Bob choosing their favored location " +
"given preference {} and recursion depth {}: {}"
.format(shared_preference, bob_depth, bob_prob))
def Marginal(fn):
return memoize(lambda *args: HashingMarginal(Search(fn).run(*args)))
def shawarma(*args, **kwargs):
return HashingMarginal(
Search(fn, max_tries=int(1e6)).run(*args, **kwargs))
