Daniel Levy
@daniellevy__
Followers
12,839
Following
334
Media
10
Statuses
767
Explore trending content on Musk Viewer
Milton
• 998115 Tweets
Florida
• 902770 Tweets
DeSantis
• 271319 Tweets
60 Minutes
• 249736 Tweets
#ファミマのスイーツパン誕生
• 123763 Tweets
Chiefs
• 111678 Tweets
Saints
• 84137 Tweets
サンデー
• 62084 Tweets
Yankees
• 40794 Tweets
#HaryanaElectionResult
• 38472 Tweets
Royals
• 38041 Tweets
Carr
• 35936 Tweets
Mahomes
• 29667 Tweets
Real Facts
• 29465 Tweets
जम्मू कश्मीर
• 28691 Tweets
Asaramji Bapu Case
• 25813 Tweets
Kelce
• 25311 Tweets
Glock
• 24440 Tweets
物価目標
• 20113 Tweets
Olave
• 13973 Tweets
Beyond excited to be starting this company with Ilya and DG! I can't imagine working on anything else at this point in human history. If you feel the same and want to work in a small, cracked, high-trust team that will produce miracles, please reach out.
Superintelligence is within reach.
Building safe superintelligence (SSI) is the most important technical problem of our time.
We've started the world’s first straight-shot SSI lab, with one goal and one product: a safe superintelligence.
It’s called Safe Superintelligence
1K
2K
14K
101
110
2K
Time to get to work ⛰️👷♂️
SSI is building a straight shot to safe superintelligence.
We’ve raised $1B from NFDG, a16z, Sequoia, DST Global, and SV Angel.
We’re hiring:
383
655
7K
46
21
899
Our paper (with Matt Hoffman and
@jaschasd
) on learning MCMC kernels parameterized by neural networks was accepted to ICLR. Up to 106x ESS and improved posterior sampling for deep generative models.
Paper:
Code:
cc:
@GoogleBrain
0
44
143
First-order methods are great but we never know which one to use. In our paper (w/ John Duchi) , we tell you which one to use and when, depending on the quadratic convexity of the constraints. Also, adaptivity matters.
To appear at
#NeurIPS2019
as an oral.
0
27
126
Two papers on optimization with differential privacy accepted to
#NeurIPS2021
!
In the 1st, with
@SZiteng
and collaborators at Google, we provide optimal algorithms to learn with *user-level* DP; a more stringent notion that protects a user's entire contribution (of many samples)
1
13
58
DRO: no-one knows what it does and it doesn't scale anyway... Or does it?
In our
#NeurIPS2020
paper, we propose optimization algorithms with running time independent of dimension and dataset size (think SGD for ERM) for CVaR and chi-square objectives. 1/4
1
18
63
On my way to
#ICLR2018
! Will be presenting our work (with Matt Hoffman and
@jaschasd
) on generalizing HMC with neural networks. Come to the poster on Thursday!
Code:
Paper:
cc:
@GoogleBrain
0
15
56
Just wrote a piece breaking down the extravagant toplaner
@FnaticsOAZ
and his raw talent!
http://t.co/UHIgqy0WrS
http://t.co/sd4i9SSPCJ
2
12
47
Choosing the optimal gradient algorithm depending on the geometry is important. Interestingly, how to do it is in seminal results about the Gaussian Sequence model. Come to our oral at 4:50pm in West Exhibition Hall A!
#NeurIPS2019
First-order methods are great but we never know which one to use. In our paper (w/ John Duchi) , we tell you which one to use and when, depending on the quadratic convexity of the constraints. Also, adaptivity matters.
To appear at
#NeurIPS2019
as an oral.
0
27
126
0
12
46
New paper out on learning under *user*-level differential privacy constraints!
In the standard DP setting, we implicitly assume that each user contributes a single sample but it turns out we often contribute many many samples (like all of our texts). 1/
1
17
32
The poster everyone has been waiting for will happen tomorrow (Wednesday) between 9-11am PST.
Come say hi and learn the secret to robustness at scale!
Paper:
Gathertown:
DRO: no-one knows what it does and it doesn't scale anyway... Or does it?
In our
#NeurIPS2020
paper, we propose optimization algorithms with running time independent of dimension and dataset size (think SGD for ERM) for CVaR and chi-square objectives. 1/4
1
18
63
0
6
25
Great collaboration with
@violet_zct
et al. on DRO for multilingual translation! Two key ideas:
- surprisingly, the robust objective improves performance on *every* language pair (vs ERM)
- tailoring the optimization algorithm to the architecture (here transformers) matters a lot
I am excited to introduce our EMNLP paper ``Distributionally Robust Multilingual Machine Translation”. To encourage uniform performance across languages, we propose a new learning objective for multilingual training based on the concept of distributionally robust optimization.
5
14
96
0
3
12
If I were to restart a PhD, I'd do it with Aditya.
I am hiring PhD students in ML this cycle! Topics include unsupervised ML, generative models, seq. decision making, and climate change.
For those interested, plz apply to UCLA CS. Deadline: Dec 15. Plz share widely!
More info:
#AI
#ML
#AcademicTwitter
4
85
422
2
1
11
@EdwardCarryLoL
Hey Eddy! Just wrote a piece on
@AlexIchLoL
and I would love your opinion; I am a fan!
http://t.co/hsp9nI5xLn
0
3
10
I've noticed that during talks a problem is never just non-convex it's HIGHLY NON-CONVEX.
1
0
8
Reviewing some heartbreaks of 2014: Part I ft.
@mithyxd
@nukeducklol
@TheSvenskeren
@AlexIchLoL
@EdwardCarryLoL
http://t.co/h6NNmtmiUp
0
5
6
@hardmaru
There is a proof of this result for arbitrary decoder and linear encoder in the very cool of
@adityagrover_
and
@ermonste
(in the case beta=0): !
0
0
7
I’ve never met a single person who has ever answered a poll. Who is answering all these polls?!
#Election2020
2
0
6
Learning with User-Level DP:
Adapting to Function Difficulty and Growth Conditions in Private Optimization:
0
0
5
Joint work with Yair Carmon, John Duchi and Aaron Sidford!
Paper:
Code:
4/4
0
0
5
I’m confused I thought AI would get us out of this pandemic???
Can you detect COVID-19 using Machine Learning? 🤔
You have an X-ray or CT scan and the task is to detect if the patient has COVID-19 or not. Sounds doable, right?
None of the 415 ML papers published on the subject in 2020 was usable. Not a single one!
Let's see why 👇
65
1K
3K
0
0
4
In the 2nd with
@HilalAsi2
and John Duchi, we show that we can *beat* the "unavoidable" dimension factor when the function has growth around the optimum.
Furthermore, we don't need to know the growth constant a priori and adapt to achieve the best rate over these families.
2
0
4
0
0
4
Some open questions remain:
- There is still a gap between the (known) lower and upper bounds for the chi-square uncertainty sets.
- What's the effect of DRO-type training for big deep learning models? This couldn't really be evaluated before!
3/4
1
0
3
2
0
3
.
@NotairesdeParis
@Notaires_CSN
y aura-t-il des mesures disciplinaires pour Maître Gildas Le Gonidec de Kerhalic?
#notaires
#touspourris
Un notaire ne peut facturer un acte inutile à 186 200 euros 😬 Cass. 1re civ., 5 juin 2019, n° 18-16.831. Lire en ligne :
5
26
48
1
2
3
Take me back
Things that didn’t exist on Thanksgiving 10 years ago:
Snapchat
Facebook Messenger
Apple Maps
Uber Eats
TikTok
Twitch
Zoom
Lyft
Slack
Peloton
Ethereum
AirPods
Apple Watch
Google Classroom
Microsoft Teams
Alexa
Tinder
Telegram
Oculus
Tesla Model S
Apple Music
Chromebook
Minecraft
44
288
2K
0
0
2
@thegautamkamath
This was a billion startup idea I've had for a while. The price of posters in the US is insane (~$100 per poster) and it is paid for every half day of conference at least. Also you can re-use TVs across conferences! Genius.
1
0
2
We propose two types of algorithm, one based on minimizing a surrogate objective and the other utilizing a more sophisticated (multilevel) gradient estimator. We show that the latter are optimal in two out of three settings and all are implemented in a few lines of PyTorch. 2/4
1
0
2
@Ramz6Up
Dude we’re trying but we just don’t have the budget this year... keep cheering for the team!!
1
0
2
@parapluieart
@Realeugenelevy
@SchittsCreek
Thank you for your support! It means a lot to us in the entertainment industry.
0
0
2
@Am1nLoL
Hey, I am a young entrepreneur who wants to start his own team. How do u think I can get in touch with players like yourself? Thx!
0
0
1
Thank you for the shoutout! cc:
@jaschasd
Generalizing Hamiltonian Monte Carlo with Neural Networks. Cool work from my colleagues
@GoogleBrain
where they train MCMC kernels parameterized by neural nets. Should improve protein folding and physics simulations.
paper:
code:
3
89
285
0
0
1
@RNGMonteCristo
@ggDoA
@PapaSmithy
Thanks a lot for the photo! It was awesome meeting you all!
http://t.co/zvbhpEupvd
0
0
2
@adityagrover_
@poolio
@Stanford
@CS_UCLA
@UCLAengineering
@facebookai
@pabbeel
@berkeley_ai
even a live (socially distanced) BPL!
1
0
2
@alexsablay
on y va?
▶️ Jean-Noël Jeanneney annonce le thème des
#RVH2021
: LE TRAVAIL.
Rendez-vous du 6 au 10 octobre 2021. 🪓
12
241
372
0
0
1
@CabochardLoL
C'est faisable man! C'est le moment de faire ta meilleure imitation de Flame et de hardcarry! J'ai foi!
0
0
1
Cuteness overloaded 😻
📽️ig: lorettabritishcattery
7
434
2K
0
0
1
1
0
1
@czgate
Thank you for the suggestion, will definitely give it a try. His themes seem to be pretty similar to Philippe Muray so I was def interested!
0
0
1
@Mridgyy
I’ve been wanting to start using roam but the lack of LaTeX macros is a bit of a road block for me (to write proofs and what not). Has it been an issue for you?
0
0
1
@Maitre_Eolas
Maître Gildas Le Gonidec de Kerhalic peut-il aussi être condamné sur le plan disciplinaire après une telle décision de la cour de cassation ?
0
0
1
@GeniMcMahon
@SchittsCreek
So glad that you're enjoying the show!! Wait until you see what we have in store for next season :)
0
0
1
@thesasho
The mnemonic for me is that Laurent Schwartz has a "t" and that Cauchy-Schwarz is the opposite.
1
0
1
@Thooorin
Just wrote a piece on AlexIch and as you are one of his admirers I would really fancy your opinion. Thanks!
http://t.co/knH8VR3P90
0
0
1
@admercs
@dustinvtran
I am not currently affiliated with Google so I cannot know for sure but I would be pretty excited by this addition to the amazing TFP!
1
0
1
1
0
1
A successful person never loses -- they either win or learn. – John Calipari
5
322
1K
0
0
1
@Cheninix
@BasedGodNorthy
Not true she has loads. Been wondering the same thing about Taylor Swift.
0
0
1
@Locodoco
@_aomen7
@samiiybl
@kithleen
@9thwonderkr
@lvl100ashe
what is the restaurant it looks great!
1
0
1
Super happy to have taken a picture with
@sjokz
! When are you leaving Paris ?
http://t.co/ClNwiTBuHI
0
0
1
@Thooorin
@RLewisReports
Really liked the bombs u dropped in Unfiltered. Can't believe nobody's mentioning it & the show has only 15k views.
1
0
1
We provide algorithms and lower bounds for these tasks (and sometimes they even match!) based on a novel generic mean estimator with error scaling as the concentration radius---think sub-Gaussian parameter---rather than the whole range of the random variable. 4/
1
0
1
I've got seven beautiful grandchildren and I will be damned if I leave them a world which is not habitable.
223
3K
9K
0
0
1
How is
@ggCMonteCristo
going to be able to do his Countdown with only western games this week?!
0
0
1
It’s a bit annoying when people attack
@NateSilver538
“90% chance for Biden to win” on the basis of it being a close election. If you win a close game a 100% of the time, 100/0 was the right forecast, no matter how close the game was.
2
0
1
0
0
1
@RNGMonteCristo
@FORG1VENGRE
@Thooorin
What better way of spending time-off than chatting with Thorin for 4 hours??
0
0
1
Reviewing some heartbreaks of 2014: Part II ft.
@CurseXpecial
@MarcelFeldkamp
@FORG1VEN247
http://t.co/G8EWwAoYeV
http://t.co/9CtAYWT4e2
0
0
1
@czgate
which one do you recommend starting with? I tried psychopolitics it's really hard for a casual reader 😱
1
0
1
@VeteranEU
Drafting is great in NA, Rush thanks to his pick/ban strategy, managed to get Elise every game!
0
1
1
@baaadas
@YSongStanford
I feel in most cases, universities might not even provide a physical, in-person option...
0
0
1
0
0
1
0
0
1
0
0
1
@shatterfront
I don't get this argument ; is it that clear that if Hillary / anyone else were president, there would be like 0 COVID deaths? The death rates in Western Europe with socialized medicine and reasonable leaders are essentially the same as in the US.
0
0
1
Gradient descent is inefficient to find saddle point (Nash equilibrium) for min-max games, because of spiralling behaviour. Beware when training your GANs …
6
126
327
1
0
1
@thejonullman
@HilalAsi2
I was being a bit facetious, apologies! I meant that it significantly improves the exponent of the d factor, e.g. if we have 5/4-growth, we go from d/n\eps -> (d/n\eps)^5! We also hypothesize that sharp growth could give rates logarithmic in dimension but haven't proved it yet 🙃
1
0
1
@JohnCLangford
@SebastienBubeck
Are you only looking for people willing to go to NYC or are you also considering remote / other locations?
0
0
1