Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Daniel Roy Profile picture
Daniel Roy
@roydanroy
Scientist. Asst Prof of Stats/CS @UofT; Founding Member @VectorInst; Canada CIFAR AI Chair. I study inference / computation to understand machine intelligence.
Kristian Georgiev Profile picture 1 subscribed
Dec 8, 2020 • 5 tweets • 2 min read
I'm concerned that my students haven't been getting the proper NeurIPS experience, since #NeurIPS2020 is online. Post a GIF that you think captures the authentic NeurIPS experience. ;) I'll start.
Sep 16, 2020 • 8 tweets • 4 min read
La Cloche Silhouette Trail complete: 78km(+) and 8 days. Made our 5 day Banff trip two years ago seem easy in comparison. Kaius (5) walked maybe 50%. Liepa (3) maybe 10%. Looking for a masseuse now with @KDziugaite. Image Kids ate way more than we anticipated. We ended up one meal short and needed way more trail snacks. Still @KDziugaite and my packs started off at 16kg and 18kg on day 1!! Yikes. Then add on our son (~21kg) and daughter (~16kg). Carrying this weight was a lot of work. Image
Aug 7, 2020 • 5 tweets • 1 min read
OK. Some NeurIPS conflict-of-interest (COI) questions for everyone. 1/n Can an AC have a COI with a paper by Y proving X if they are in a race to prove X with authors Y?
Jul 16, 2020 • 5 tweets • 1 min read
How’s attending ICML with kids? Me: haven’t even registered yet, despite having three posters to present. Luckily I have excellent coauthors and my wife registered.
Jun 6, 2020 • 5 tweets • 1 min read
Wow. I just stepped out of my own filter bubble by reading an article on Brussels Times. Check out the ads on a single page. 1/n. Image
May 23, 2020 • 8 tweets • 4 min read
Information-theoretic generalization bounds go brrrrr....
arxiv.org/abs/2005.08044

Nice ideas in here combining Catoni's single draw setting and the @shortstein @zakynthinou framework. I appreciate multiple cites to my group's recent work. Image Our most recent paper extending Steinke--Zakynithinou is most relevant: arxiv.org/abs/2004.12983

Yes, we assume bounded loss, but subG bounds are trivial.
Jan 17, 2020 • 4 tweets • 1 min read
Grappling with devastating news: my four-year old has Dunning-Kruger. 🙏🙏🙏🙏 Happy to hear from other parents who could provide advice / tips. E.g., he's gone skiing a few times, but now acts like he's an expert and barks commands about skiing at me.