I keep seeing lots of long-time #fediverse users saying 'don't favourite posts it does nothing' but actually when you favourite my posts it makes me smile and I'm sure I'm not the only one.
Associate Professor of Machine Learning, University of Oxford
OATML Oxford Group Leader
http://yarin.co
I keep seeing lots of long-time #fediverse users saying 'don't favourite posts it does nothing' but actually when you favourite my posts it makes me smile and I'm sure I'm not the only one.
🎉 #Mastodon 4.0 is out now! This server software update includes a ton of improvements, like following hashtags, translating posts, editing, an improved filtering system, customizable user roles for administration, but also some important security fixes.
Check out the full changelog:
Unfollow that account you hate. Stop reading that book you think is meh. Close all those tabs.
Life is simply too short.
Some excellent math jokes in the UC workers strike picket line https://www.capradio.org/articles/2022/11/14/nearly-50000-university-of-california-graduate-student-employees-launch-open-ended-strike/
#ICML2023 deadlines are out:
🚨Paper Submissions Open on @openreviewnet:
🗓️Jan 09 '23 02:00 PM UTC
🚨Full Paper Submission Deadline:
Jan 26 '23 08:00 PM UTC
I'm a senior researcher at Microsoft Research AI4Science in Amsterdam.
My research interests include AI4Science, single- and multi-agent reinforcement learning, and structure, symmetry, and equivariance in deep learning.
#reinforcementlearning #ai4science #machinelearning #deeplearning #equivariantagents
*mic check! one two! one two!
here's my genesis block #introduction
I'm a PhD student working with @yaringal in OATML at the university of Oxford.
I'm working in bayesian sequential decision-making, experimental design and causality. I'm also interested in generative design and architecture, artificial life, cybernetics, biology and music. More about me here https://ptigas.com/
Looking back on the last decade, it seems clear that Twitter caused some pretty significant shifts in how ML/AI research was driven. Papers could go viral in a way that wasn't really possible at conferences or on arXiv. This incentivized lots of work in areas that could grab your attention in a tweet - anything with cool videos or visuals. The current wave of generative AI probably wouldn't exist if it wasn't for social media. Was this a good thing? Will academics moving off Twitter change that?
Developing Mastodon, I strive to use terminology that is familiar to as many people as possible so as to not put up unnecessary barriers in understanding. For this reason I am happy that we went away from calling posts "toots". It has always been a point of friction for people. Not that I dislike branding, quite the opposite, but in my experience receiving feedback over half a decade, it had a negative effect on the platform for little benefit.
@mathieualain @Deisenroth @yaringal @emtiyaz @imurray @naesseth @misovalko @nicolas_le_roux @pfau @jwvdm @sirbayes @avehtari @rockt @elisevanderpol @nantas @timnitGebru @msalbergo @Inference @negar @davidbarber strangely disappointed I wasn’t in the Bayes section. Weird since I spent like a decade trying to distance myself from that. I guess I’m back?
3️⃣ out of 9️⃣
BAYESIAN/GAUSSIAN PROCESSES:
@deisenroth (RL)
@yaringal
@davidbarber
@emtiyaz
@imurray
@jwvdm
@sirbayes
@avehtari
@naesseth
@Inference
@mguindani
@PhilippHennig
@timrudner
@maosbot
@jmtomczak
@markvanderwilk
@SimoSarkka
⬇️ ⬇️ ⬇️
@mat_kelcey with probability 0.5, a frozen randomly initialised a linear controller will solve cartpole..
1️⃣ out of 9️⃣
If you also had trouble finding people to follow in AI/ML/statistics, here is a partial list.
Please help me to complete and share this list!
PS: I have tried to put people into subcategories. Don't be offended if you feel you are not in the right one (and just tell me if it bothers you).
PS PS: I sometimes have to move the names around to accommodate the character limit. I'm sorry if that sends notifications. If you want to be removed please DM me.
⬇️ ⬇️ ⬇️
Hi! I’m Vicki. I work as a machine learning engineer. My main interests are machine learning in prod and software best practices, distributed systems, recommender systems, Python, the JVM, and getting CI to build.
I also am very interested in how humans and tech work together and sometimes write a newsletter about it, Normcore Tech.
When I’m not rerunning black, I am caring for two small children and planning #NormConf.
I love Twitter but I’m here now too.
#introduction
I am an astronomer at Helsinki University, studying exoplanets and active stellar surfaces.
Among the dozens of planets I have discovered, is the nearest candidate habitable world Proxima b, orbiting the very nearest star to the Sun.
The discovery didn't happen in an instance. It was a process that began almost a decade ago. https://physicstoday.scitation.org/do/10.1063/pt.5.2048/full/
The server journa.host for working journalists is picking up a lot of interesting users with a lot of publishing prowess and name recognition. Feels like watching the birth of a new media center.
https://journa.host