Boris LEISTEDT

Really clever way to start the #AstroHackWeek zoom meeting with some down time for people to set up their audio and also annotate the starting slide for some fun.

Second day of #AstroHackWeek! Focus to day is on all the hacking :D

We know many of you have been asking about the progress we’ve made to look into our history and legacy and today we can update you (THREAD 🧵) More info👇🏽 ow.ly/6tpP50AmKkA

Awesome wrap up session at #AstroHackWeek ! So many great hacks by a diverse group of lovely people. Best week of the year again

Popular breakout in types of neural networks, architectures, applications advice, etc #AstroHackWeek

Breakout on mcmc and sampling methods yesterday! #AstroHackWeek

Second day at #AstroHackWeek! Bayesian inference tutorials in the morning, quantum computing breakout and hacking in the afternoon!

Hacking explained #AstroHackWeek

Awesome data visualization tutorial at #AstroHackWeek

Astro Hack Week 2018 is happening! Time for a Tea Party team building exercise #AstroHackWeek

Data visualization tutorial at #AstroHackWeek

Applications for #AstroHackWeek are now open (until April 30)! We invite researchers of all backgrounds/career stages interested in astronomical data analysis, to come work on exciting projects with us! Details + application here: astrohackweek.org/2018/ #astronomy #DataScience

Thrilled to announce that Astro Hack Week 2018 will be taking place on August 6-10 at the Lorentz Center in Leiden, the Netherlands. Mark your calendars! #astrohackweek

Oh, astrobite did a piece about our paper! cc @Galactichawk @davidwhogg @jobovy astrobites.org/2017/11/30/gai…

@KyleCranmer @dustinvtran @NYUDataScience @NYUScience I love Edward! (Looking forward to seeing it in TF). None of those projects used it, but some of our new (related) work does

Slides of the colloquia I gave in UMichigan and UofToronto. Wonderful visits, people, and research! I'm so excited now speakerdeck.com/ixkael/data-dr…

Getting ready for Gaia DR2: multicolor CMD-GMM working (with tensorflow SGD), 3D dust is next! cc @davidwhogg

Analytic marginalization of additive and multiplicative scalings in gaussian likelihoods goo.gl/tYRJon cc @davidwhogg

@davidwhogg MAP of the GMM-CMD, via nested sampling or rejection+optimization. But for multicolors and DR2+ let's use stochastic gradients+tensorflow!

Improved CMD-GMM model with parallaxes robustly marginalized. Now adding more colors & 3D dust + tensorflow for Gaia DR2+. cc @davidwhogg