def variational_objective(params, t):
"""Provides a stochastic estimate of the variational lower bound."""
mean, log_std = unpack_params(params)
samples = rs.randn(num_samples, D) * np.exp(log_std) + mean
lower_bound = gaussian_entropy(log_std) + np.mean(logprob(samples, t))
loss = np.mean(logprob(samples, t))
print("loss is "+ str(loss))
return -lower_bound
def variational_objective(params, t):
"""Provides a stochastic estimate of the variational lower bound."""
mean, log_std = unpack_params(params)
generatedSample=rs.randn(num_samples, D) * np.exp(log_std)
samples = generatedSample + mean
#samples: sample of weights
#t: targets
#inputs used in logprob is the inputs user initial generated
logvalue = logprob(samples, t)
lower_bound = gaussian_entropy(log_std) + np.mean(logprob(samples, t))
loss = np.mean(logprob(samples, t))
print("loss is "+ str(loss))
return -lower_bound
def fit_maxlike(x, r_guess):
# follows Wikipedia's section on negative binomial max likelihood
assert np.var(x) > np.mean(x), "Likelihood-maximizing parameters don't exist!"
loglike = lambda r, p: np.sum(negbin_loglike(r, p, x))
p = lambda r: np.sum(x) / np.sum(r+x)
rprime = lambda r: grad(loglike)(r, p(r))
r = newton(rprime, r_guess)
return r, p(r)
def lift(vy): return np.mean(vy - init_vy)
def objective(params):
def drag(vx): return np.mean(init_vx - vx) def lift(vy): return np.mean(vy - init_vy)
def frac_err(W_vect, X, T):
return np.mean(np.argmax(T, axis=1) != np.argmax(pred_fun(W_vect, X), axis=1))
def distance_from_target_image(smoke):
return np.mean((target - smoke)**2)