Clément Canonne
@ccanonne.github.io
Senior Lecturer #USydCompSci at the University of Sydney. Postdocs IBM Research and Stanford; PhD at Columbia. Converts ☕ into puns: sometimes theorems. He/him.
There are four kangaroos in this picture. #recursion
November 9, 2025 at 8:13 AM
There are four kangaroos in this picture. #recursion
Spotting a Canberra kangaroo on its way to Parliament
November 9, 2025 at 7:31 AM
Spotting a Canberra kangaroo on its way to Parliament
Canberra is such a pleasure to visit.
November 9, 2025 at 6:15 AM
Canberra is such a pleasure to visit.
Very good design, no notes.
November 8, 2025 at 12:39 AM
Very good design, no notes.
You monsters, WHAT HAVE YOU DONE TO RUDOLPH?!
November 7, 2025 at 11:08 AM
You monsters, WHAT HAVE YOU DONE TO RUDOLPH?!
Saw this fluffball in the neighborhood coffee shop.
November 2, 2025 at 11:04 AM
Saw this fluffball in the neighborhood coffee shop.
Exciting! Fingers crossed now... "The petition has been presented to the House and is awaiting a response from the Minister responsible for the matter(s) raised in the petition request." 🇦🇺
www.aph.gov.au/e-petitions/...
www.aph.gov.au/e-petitions/...
October 27, 2025 at 9:53 AM
Exciting! Fingers crossed now... "The petition has been presented to the House and is awaiting a response from the Minister responsible for the matter(s) raised in the petition request." 🇦🇺
www.aph.gov.au/e-petitions/...
www.aph.gov.au/e-petitions/...
This @smbccomics.bsky.social on quantum computing really hits the mark regarding teaching theoretical computer science more broadly―or, I reckon, science.*
www.smbc-comics.com/comic/intuit... by @zachweinersmith.bsky.social
*Hey, how would I know, I'm not a scientist.
www.smbc-comics.com/comic/intuit... by @zachweinersmith.bsky.social
*Hey, how would I know, I'm not a scientist.
October 26, 2025 at 10:39 PM
This @smbccomics.bsky.social on quantum computing really hits the mark regarding teaching theoretical computer science more broadly―or, I reckon, science.*
www.smbc-comics.com/comic/intuit... by @zachweinersmith.bsky.social
*Hey, how would I know, I'm not a scientist.
www.smbc-comics.com/comic/intuit... by @zachweinersmith.bsky.social
*Hey, how would I know, I'm not a scientist.
It's purple tree season again!
October 25, 2025 at 11:10 PM
It's purple tree season again!
User-level privacy means that the local DP guarantee holds w.r.t. an entire user: the algo's output shouldn't leak much about any given user, even if their *all* of their m samples changed!
This is much harder than the m=1 case... and we want *practical* algos: no shared random seed, for instance!
This is much harder than the m=1 case... and we want *practical* algos: no shared random seed, for instance!
October 22, 2025 at 4:18 AM
User-level privacy means that the local DP guarantee holds w.r.t. an entire user: the algo's output shouldn't leak much about any given user, even if their *all* of their m samples changed!
This is much harder than the m=1 case... and we want *practical* algos: no shared random seed, for instance!
This is much harder than the m=1 case... and we want *practical* algos: no shared random seed, for instance!
David Harvey (UNSW) giving a very engaging talk at our Sydney Algorithms and Computing Theory (SACT) seminar on his upcoming #FOCS2025 paper,
"Integer multiplication is at least as hard as matrix transposition." @sydneycompsci.bsky.social
"Integer multiplication is at least as hard as matrix transposition." @sydneycompsci.bsky.social
October 22, 2025 at 2:44 AM
David Harvey (UNSW) giving a very engaging talk at our Sydney Algorithms and Computing Theory (SACT) seminar on his upcoming #FOCS2025 paper,
"Integer multiplication is at least as hard as matrix transposition." @sydneycompsci.bsky.social
"Integer multiplication is at least as hard as matrix transposition." @sydneycompsci.bsky.social
In case you're wondering
October 20, 2025 at 7:32 AM
In case you're wondering
Dec 11-13 at #USyd, just before #FOCS2025: "A Celebration of Theoretical Computer Science": now with a poster!
Info and (free) registration: sites.google.com/view/celebra... #TCSSky
Info and (free) registration: sites.google.com/view/celebra... #TCSSky
October 20, 2025 at 3:29 AM
Dec 11-13 at #USyd, just before #FOCS2025: "A Celebration of Theoretical Computer Science": now with a poster!
Info and (free) registration: sites.google.com/view/celebra... #TCSSky
Info and (free) registration: sites.google.com/view/celebra... #TCSSky
The potato is still growing. All hail our new potato master. 🥔
October 18, 2025 at 5:53 AM
The potato is still growing. All hail our new potato master. 🥔
"Graphic design is my passion" (creating a logo/poster)
October 13, 2025 at 10:21 AM
"Graphic design is my passion" (creating a logo/poster)
Truly a fantastic line-up of speakers! Also: poster session (most likely involving gelato), reception, networking, Q&A: sites.google.com/view/celebra...
All that on the (beautiful) #USyd campus, in summery December, just before FOCS! #TCSSky
All that on the (beautiful) #USyd campus, in summery December, just before FOCS! #TCSSky
October 12, 2025 at 3:13 AM
Truly a fantastic line-up of speakers! Also: poster session (most likely involving gelato), reception, networking, Q&A: sites.google.com/view/celebra...
All that on the (beautiful) #USyd campus, in summery December, just before FOCS! #TCSSky
All that on the (beautiful) #USyd campus, in summery December, just before FOCS! #TCSSky
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.)
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 $0=a-a$. 🤷)](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:ac6qioenkpl3jmmnib3tpbhb/bafkreieevv3petzx4kqxx3o6i4oqtugsjh52n7i2ellidxmetefboag6fi@jpeg)
October 11, 2025 at 4:04 AM
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.)
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.)
I saw one of your minions in my street this morning. In my STREET.
October 11, 2025 at 1:17 AM
I saw one of your minions in my street this morning. In my STREET.
I... I just don't know what to do
October 3, 2025 at 6:23 AM
I... I just don't know what to do
October 3, 2025 at 5:47 AM
Talk featuring, very topically, the platypus channel. 🦫 🦆
September 30, 2025 at 11:23 PM
Talk featuring, very topically, the platypus channel. 🦫 🦆