🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Refreshing lecture by David Dunson on "Machine learning for scientific inferences: Debunking the hype". Interested to see that work to remove bias from ML is making progress thanks to Kristian Lum and James Johndrow, Ph.D. arxiv.org/abs/1703.04957

Publishing your paper in a Taylor and Francis journal that even Stanford apparently doesn't pay for and also not putting a manuscript on arXiv is a great way to make sure your work has zero impact.

I wish you’d called us Harvard Humanitarian Initiative @HHI_Signal first. This isn’t what the disaster relief field needs right now. In fact, sector needs investments in how to use tech we’ve got/use tech less. Most pressing challenges are all about political will, protection, human rights, not more code x.com/developerworks…

Google discovered that giving every employee a credit card and trusting them to follow policy on expenses, even with some bad actors, was cheaper than the enforcement and structures most companies use. How much do you waste due to low trust? Can you do the math for your company?

wow!!!!!! love it. that is out of the box thinking. congrats Atul Gawande x.com/jtozz/status/1…

Funny but not so true anymore. Post selective inference is largely figured out for forward stepwise, see statweb.stanford.edu/~tibs/ftp/spac…. Of course causal inference always requires something more but it seems odd to pick on forward stepwise given recent developments. x.com/edwardtufte/st…

This headline from The Hill is embarrassing. I get that they're quoting someone but I thought part of the point of journalism is not to publish things that are wrong. Headline should read "pollster doesn't understand probability, blames FiveThirtyEight" x.com/NateSilver538/…

Getting excited for #jsm2018. Come say hi if you want to talk about algorithmic fairness, stats for human rights, (gender-based) harassment in our field, OR if you want to tell me about your own research! I’d love to hear about it.

a whole publishing industry has emerged to assuage boomer guilt for having robbed their children

Great crowd here at #jsm2018 at the session on interpretable ML models.... and we snuck in a talk on fairness. 😁 If you’re interested in this topic, consider submitting to/attending FAT*! x.com/fatconference/…

Excited to be on a "brilliant women in ai ethics" list with so many friends and women that I think are brilliant! @timnitGebru (@dair-community.social/bsky.social) Meredith Whittaker Kate Crawford Dr. Joy Buolamwini Joanna J Bryson Karen Levy Kate Darling @kldivergence @ShannonVallor Terah Lyons AI4ALL Internet Ethics medium.com/@miad/100-bril…

Snow day today; called it yesterday evening based on predictive modeling. I’d like to invite the people who convinced the world that Machine Learning has solved predictive problems to come over and hang out w/ my four very hyper & disappointed kids. #stopthehype #keeptheirtrust

Robin Gomila 1/ I think the sentence: "I draw on econometric theory and established statistical findings to demonstrate that linear regression (OLS) is generally the best strategy to estimate causal effects on binary outcomes" will make some economists rethink if "Almost Harmless" is truly

Jurgen Vinju mireillemoret Kristian Lum has a good take on this. I want to maybe help clarify a little bit here. It sounds like Jurgen Vinju (and maybe mireillemoret?) is talking about hypothesis testing...

Critical issue right now. "Get it yourself" will lead to disaster.

We’ve been working on a number of projects with interdisciplinary teams of people at Penn including at the hospital. Check out the latest.

Turns out you can get acceptable linear algebra performance in #rstats if you build from source. Still much slower than #julialang but at least not orders of magnitude slower.

Definitely critical if you want to do fast numerical linear algebra. Now no one has an excuse for not having mkl. Thanks for doing this Dirk Eddelbuettel ! x.com/eddelbuettel/s…

Proud of Kristian Lum for her profile in Bloomberg today. bloomberg.com/news/articles/…

I guess I'm just saying I don't think this is the prevailing view among statisticians. Do you? x.com/drob/status/99…