626_privacy/tensorflow_privacy
2
0
Fork 1
Code Issues Pull requests 1 Projects Releases Packages Wiki Activity Actions

Fix numerical instability in computing A(alpha) for very large integer alpha.

Browse source 
Tested that new implementation agrees with existing implementation on all small integers but also scales to 10^6.

PiperOrigin-RevId: 348492489
This commit is contained in:
Galen Andrew 2020-12-21 10:51:46 -08:00 committed by A. Unique TensorFlower
parent 276d2d74d5
commit e4f9794542
1 changed files with 6 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
tensorflow_privacy/privacy/analysis/rdp_accountant.py
View file

@ -85,6 +85,11 @@ def _log_print(logx):
return "exp({})".format(logx)
def _log_comb(n, k):
return (special.gammaln(n + 1) -
special.gammaln(k + 1) - special.gammaln(n - k + 1))
def _compute_log_a_int(q, sigma, alpha):
"""Compute log(A_alpha) for integer alpha. 0 < q < 1."""
assert isinstance(alpha, six.integer_types)
@ -94,8 +99,7 @@ def _compute_log_a_int(q, sigma, alpha):
for i in range(alpha + 1):
log_coef_i = (
math.log(special.binom(alpha, i)) + i * math.log(q) +
(alpha - i) * math.log(1 - q))
_log_comb(alpha, i) + i * math.log(q) + (alpha - i) * math.log(1 - q))
s = log_coef_i + (i * i - i) / (2 * (sigma**2))
log_a = _log_add(log_a, s)