aboutsummaryrefslogtreecommitdiff
path: root/posts
diff options
context:
space:
mode:
authorYuchen Pei <me@ypei.me>2019-03-19 17:12:26 +0100
committerYuchen Pei <me@ypei.me>2019-03-19 17:12:26 +0100
commit0a8781825217d4e3b15ea757142eb957dc02d9ec (patch)
treef52d3b7f32801f2aedf3d2e41bc499e4c68c0d25 /posts
parentac8979f6770d2c1ed79b680fc720cbefd8d588c7 (diff)
updated due to conjecture 1 probably false
Diffstat (limited to 'posts')
-rw-r--r--posts/2019-03-14-great-but-manageable-expectations.md14
1 files changed, 10 insertions, 4 deletions
diff --git a/posts/2019-03-14-great-but-manageable-expectations.md b/posts/2019-03-14-great-but-manageable-expectations.md
index d622fd4..d5ca684 100644
--- a/posts/2019-03-14-great-but-manageable-expectations.md
+++ b/posts/2019-03-14-great-but-manageable-expectations.md
@@ -341,7 +341,7 @@ then we have $(\lambda, O(r^2 \lambda / \sigma^2))$-rdp.
Wait, why is there a conjecture? Well, I have tried but not been able to
prove the following, which is a hidden assumption in the original proof:
-**Conjecture 1**. Let $p_i$, $q_i$, $\mu_i$, $\nu_i$ be
+**Conjecture 1** \[Probably [FALSE](https://math.stackexchange.com/a/3152296/149540), to be removed\]. Let $p_i$, $q_i$, $\mu_i$, $\nu_i$ be
probability densities on the same space for $i = 1 : n$. If
$D_\lambda(p_i || q_i) \le D_\lambda(\mu_i || \nu_i)$ for all $i$, then
@@ -351,9 +351,12 @@ Basically, it is saying \"if for each $i$, $p_i$ and $q_i$ are closer to
each other than $\mu_i$ and $\nu_i$, then so are their averages over
$i$\".
So it is heuristically reasonable.
+But it is probably [**FALSE**](https://math.stackexchange.com/a/3152296/149540).
+<!---
This conjecture is equivalent to its special case when $n = 2$ by an induction argument
(replacing one pair of densities at a time).
+-->
Recall the definition of $G_\lambda$ under the definition of Rényi
differential privacy. The following Claim will be useful.
@@ -381,7 +384,9 @@ I will break the proof into two parts:
$c_1$, $c_2$ and the function $C(c_1, c_2)$ are important to the
practicality and usefulness of Conjecture 0.
-Part 1 can be derived using Conjecture 1.
+Part 1 can be derived using Conjecture 1, but since Conjecture 1 is probably false,
+let us rename Part 1 itself _Conjecture 2_, which needs to be verified by other means.
+<!---
We use the notations $p_I$ and $q_I$ to be $q$ and $p$ conditioned on
the subsampling index $I$, just like in the proof of the subsampling theorems (Claim 19 and 24).
Then
@@ -400,6 +405,7 @@ $|\mathcal I_\in| = r |\mathcal I|$, by Conjecture 1, we have Part 1.
it suffices to prove a weaker version of it, by specialising on mixture of Gaussians,
in order to have a Claim 26 without any conjectural assumptions.
I have in fact posted the Conjecture on [Stackexchange](https://math.stackexchange.com/questions/3147963/an-inequality-related-to-the-renyi-divergence).
+-->
Now let us verify Part 2.
@@ -486,7 +492,7 @@ In the following for consistency we retain $k$ as the number of epochs,
and use $T := k / r$ to denote the number of compositions / steps /
minibatches. With Conjecture 0 we have:
-**Claim 28**. Assuming Conjecture 1 is true. Let $\epsilon, c_1, c_2 > 0$,
+**Claim 28**. Assuming Conjecture 2 is true. Let $\epsilon, c_1, c_2 > 0$,
$r \le c_1 \sigma^{-1}$,
$T = {c_2 \over 2 C(c_1, c_2)} \epsilon \sigma^2$. then the DP-SGD with
subsampling rate $r$, and $T$ steps is $(\epsilon, \delta)$-dp for
@@ -733,7 +739,7 @@ Here is a list of what I think may be interesting topics or potential
problems to look at, with no guarantee that they are all awesome
untouched research problems:
-1. Prove Conjecture 1
+1. Prove Conjecture 2
2. Find a theoretically definitive answer whether the methods in Part 1
or Part 2 yield better privacy guarantees.
3. Study the non-Gaussian cases, general or specific. Let $p$ be some