Clément Canonne
@ccanonne.github.io
6.7K followers 560 following 2.3K posts
Senior Lecturer #USydCompSci at the University of Sydney. Postdocs IBM Research and Stanford; PhD at Columbia. Converts ☕ into puns: sometimes theorems. He/him.
Posts Media Videos Starter Packs
Pinned
ccanonne.github.io
Reminder/plug: my graduate-level monograph on "Topics and Techniques in Distribution Testing" (FnT Comm. and Inf Theory, 2022).

📖 ccanonne.github.io/survey-topic... [Latest draft+exercise solns, free]
📗 nowpublishers.com/article/Deta... [Official pub]
📝 github.com/ccanonne/sur... [LaTeX source]
Table of contents of the monograph
ccanonne.github.io
Not that I can see (for instance, it's an equality, not an inequality). At least, the connection, if any, is not obvious to me.
Reposted by Clément Canonne
mcnees.bsky.social
“Why is the night sky dark, if we live in an infinite universe?”

Kepler, Halley, and Cheseaux all pondered this apparent paradox, but the question is commonly attributed to Heinrich Olbers, who was born #OTD in 1758. 🧪 🔭

Images: hyperphysics.phy-astr.gsu.edu/hbase/Astro/..., Wellcome Collection
A diagram shows arrows radiating from a central point, indicating directions, with lots of white disks (stars) surrounding the central point. An etching of Hans Olbers. He wears a heavy coat, vest, and high collared white shirt. His wispy hair is receding.
ccanonne.github.io
"With the whole GPU craze, now the most valuable rare metal is clearly potatium.
- You mean potassium?
- No, potatium. That's what they make chips from."
Reposted by Clément Canonne
ccanonne.github.io
With the same idea, one can show that if X, Y are independent random variables (both with finite variances) then

𝔼[(X-Y)²] = Var[X]+Var[Y]+(𝔼[X]-𝔼[Y])²

which in particular establishes the "standard useful fact" that, if X and Y are i.i.d., then 𝔼[(X-Y)²]=2Var[X].
ccanonne.github.io
With the same idea, one can show that if X, Y are independent random variables (both with finite variances) then

𝔼[(X-Y)²] = Var[X]+Var[Y]+(𝔼[X]-𝔼[Y])²

which in particular establishes the "standard useful fact" that, if X and Y are i.i.d., then 𝔼[(X-Y)²]=2Var[X].
ccanonne.github.io
Here's a classic (but fun to show) fact: if X is any random variable (with a finite variance) and λ is a real, then

𝔼[(X-λ)²] = Var[X]+(𝔼[X]-λ)²

(In particular, this shows that 𝔼[X] is the quantity minimizing 𝔼[(X-λ)²] over all λ, and that Var[X] is the resulting value.)
A short proof: here is the LaTeX code.

**Proof.** We have, for any $\color{blue}{\lambda} \in\mathbb{R}$,
\begin{align*}
\mathbb{E}[(X-\color{blue}{\lambda})^2]
&= \mathbb{E}[(X-\color{red}{\mathbb{E}[X]} + \color{red}{\mathbb{E}[X]} - \color{blue} {\lambda})^2] \\
&=\mathbb{E}[(X-\color{red}{\mathbb{E}[X]})^2 + 2(X-\color{red}{\mathbb{E}[X]})(\color{red}{\mathbb{E}[X]} - \color{blue} {\lambda}) + (\color{red}{\mathbb{E}[X]} - \color{blue} {\lambda})^2]\\
&=\underbrace{\mathbb{E}[(X-\color{red}{\mathbb{E}[X]})^2]}_{=\textrm{Var}[X]} + 2\underbrace{\mathbb{E}[X-\color{red}{\mathbb{E}[X]}]}_{=0}(\color{red}{\mathbb{E}[X]} - \color{blue} {\lambda})] + (\color{red}{\mathbb{E}[X]} - \color{blue} {\lambda})^2
\end{align*}
and that's all. (The first step is a trick known as *"hiding zero:"* writing $0=a-a$. 🤷)
ccanonne.github.io
(My original plan was 1000 🤷)
ccanonne.github.io
Turns out, "I will buy and unleash 500 earthworms upon our small garden" was not a sentence my partner was keen to hear on this beautiful Saturday.
ccanonne.github.io
Yes! And good idea: we have a few in December! (FOCS, AsiaCrypt, and two workshops before and after FOCS)
ccanonne.github.io
I saw one of your minions in my street this morning. In my STREET.
A grey, white-faced heron on a brick fence, in front of a house. A close-up shot of a white-faced heron on a brick fence.
ccanonne.github.io
(just to be a little more nuanced, as the above came off curter than intended: scores are not useless, but they're only part of thr assessment process. Looking only at them, decisions may look arbitrary, but often (hopefully) it's due to extra context and discussion that the reviews provide/hint at)
ccanonne.github.io
"Misses one* of* the key points" (I missed out on a chance to spellcheck my post)
ccanonne.github.io
A poll like this misses on the key points of peer review (and one that feeds many, many wrong takes in my view): a review is not just a score, and acceptance/rejection isn't just a function of the 3 or 4 scores.

But asking to suggest a decision based on that info only reinforces that misconception!
Reposted by Clément Canonne
robinkothari.bsky.social
Fresh on the arXiv: @booleananalysis.bsky.social, Kewen Wu, and I present new classical algorithms for the Short Integer Solution problem (under infinity norm) that outperform the elegant Chen-Liu-Zhandry quantum algorithm, showing that there is no exponential quantum speed up anymore.
ccanonne.github.io
I wasn't allowed to post my latest quantum computing joke. Was told it violated the no-clowning theorem
ccanonne.github.io
Happy QIP arXiv posting deadline to all who celebrate!
Reposted by Clément Canonne
tcsplus.bsky.social
The recording and slides of this week's talk by Janani Sundaresan, on how "Distributed Triangle Detection is Hard in Few Rounds", are now available online!

www.tcsplus.org/welcome/past... #TCSSky
TCS+ - 2025-2026
2025/10/08: Janani Sundaresan, "Distributed Triangle Detection is Hard in Few Rounds" Janani Sundaresan (University of Waterloo)
www.tcsplus.org
ccanonne.github.io
Fyi, if you follow me here and only read my rants or quips about emails, workload, or lack of sleep*

Things are good. Academic life is good, or can be — up to you. It's not perfect, but it's pretty great! (As the RL people hope to say: no regret!)

*Blame the kookaburra laughing at my window at 5am
ccanonne.github.io
Oh, come on, English. I wasted HOURS trying to find a proof, because of you.

Do you even realize how close *unprovable* and *improvable* sound to me?
Reposted by Clément Canonne
aaroth.bsky.social
The FORC 2026 call for papers is out! responsiblecomputing.org/forc-2026-ca... Two reviewing cycles with two deadlines: Nov 11 and Feb 17. If you haven't been, FORC is a great venue for theoretical work in "responsible AI" --- fairness, privacy, social choice, CS&Law, explainability, etc.
FORC 2026: Call for Papers
The 7th annual Symposium on Foundations of Responsible Computing (FORC) will be held on June 3-5, 2026 at Harvard University. Brief summary for those who are familiar with past editions (prior to 2…
responsiblecomputing.org