Checkout @RaphaelS101 post about benchmarking python keras survival models from #pycox in #mlr3 that even includes automatic tuning. towardsdatascience.com/neural-network…

Nice read on overfitting through tuning in classical ML. Relevant ideas go beyond the application of Bayesian optimization. arxiv.org/pdf/2104.08166… #kdd2021 #MachineLearning

Instead of scrolling through Twitter, Insta etc. you can keep scrolling here for a while mkorostoff.github.io/1-pixel-wealth/

I wonder how many hyperparameter optimization papers are just an instance of this one line: #stats #MachineLearning

Why have I never head before of the NIPS 2014 experiment? Two committees judging the same 166 papers: "about 57% of the papers accepted by the first committee were rejected by the second one and vice versa" blog.mrtz.org/2014/12/15/the…

Information hygiene is truly the issue of our time. If we don't fix it, our ability to address other major issues -- from the climate crisis to deadly pandemics -- takes a massive hit.

Given that 90% of scientific research is about understanding the work of others and building on it, and given the v. high long-term returns of science, it's surprising how little attention is dedicated to systemically optimizing scientific communication and knowledge discovery

We considerably (!) updated our paper on hyperparameter optimization. It now version v3. Have a look. Particularly interesting for new users of HPO, who are looking for a principled overview. arxiv.org/abs/2107.05847 #MachineLearning #AutoML #HPO

Looking for an open Ph.D. position in interpretable #MachineLearning at Universität München in SLDS / Bernd Bischl's group? Want to build on the research of Christoph Molnar 🦋 christophmolnar.bsky.social? Apply now: lmu.de/en/about-lmu/w… #interpretableML #iml #ExplainableAI #xai #DataScience

Has anyone ever calculated the (environmental) cost of having the same product from so many different brands and in various types?

"There’s no scientific integrity if you can’t control for human desperation." bethanyactually.tumblr.com/post/664610567…

“Public key” is kind of an oxymoron to most peoples perception of a key. Naming it “private key” and “public lock” would probably have helped. 🤔

A colleague created this very nice and sleek weighted voting website for consensus decision-making. You might want to try it for your next group poll on where to go for dinner, or even for something more serious. lets.konsens.it #konsensit

Is there a dev workflow where you write software specifications under version control and a PR to this specification is linked to a PR in the code? It feels wrong to have specs only in tickets/tasks.

Amazing to see your papers being published after being out of academia for a while. A big thank you to all collaborators!

Join us at the #Berlin R Users Meet-Up to talk about #MachineLearning in R. Sebastian Fischer and me will present the current state of ML Pipelines in #mlr3. 📅Tue, 21th Nov 6PM 📍@spicedacademy #RStats #meetup #DataScientists