Day 45: Clique Culture

Important talk by @timnitGebru at @CVPR on avoiding "clique culture": cliques push people out, harm diversity, & we all miss out on good ideas https://t.co/Zs8Tka2nDZ

— Rachel Thomas (@math_rachel) July 4, 2018

Here are some tips from @timnitGebru on how to counter clique culture: pic.twitter.com/2RzPURReGp

— Rachel Thomas (@math_rachel) July 4, 2018

Another example of how informal social interactions can have a big impact on who succeeds is tech startups that lack onboarding https://t.co/chSF29FQ7O pic.twitter.com/9CsLPYBsmC

— Rachel Thomas (@math_rachel) July 4, 2018

This new family of GAN loss functions looks promising! I'm especially excited about Fig 4-6, where we see that the new loss results in much faster learning during the first several iterations of training. I implemented the RSGAN loss on a toy problem and it worked well. https://t.co/BgVPQZE4Nx

— Ian Goodfellow (@goodfellow_ian) July 3, 2018

In our latest paper we study coordinated hate attacks against YouTube users and try to answer the question: can we predict, at upload time, whether a video will likely attract attacks in the future? https://t.co/snhpMK1DoM pic.twitter.com/YNOPny1reU

— Gianluca Stringhini (@gianluca_string) May 22, 2018

How can one detect friendly and adversarial behaviour from raw data? https://t.co/cRJCyfo5oo

— DeepMind (@DeepMindAI) July 3, 2018

Green=more likely to die in home
Purple=more likely to die in facilities https://t.co/Hs7wq6kYLI

— MetricMaps (@MetricMaps) July 1, 2018

pic.twitter.com/JpGwNQrHiI

— FiveThirtyEight (@FiveThirtyEight) July 3, 2018

I charted all the batteries https://t.co/KEemPr1hj5 pic.twitter.com/3xN4fdPqhx

— Nathan Yau (@flowingdata) July 3, 2018

Visualizing the human footprint. #dataviz

Full set of dataviz here: https://t.co/MLB6iJ1Yk1 pic.twitter.com/KRSwnUPx5K

— Randy Olson (@randal_olson) July 3, 2018

Presenting survey data https://t.co/E9D9ROQBor #rstats #DataScience

— R-bloggers (@Rbloggers) July 3, 2018

a v. special #July4 #plotswithchristine adds a dash of "chart junk" to my usual #dataviz fun––UFO sightings peak on #IndependenceDay 👽🛸👾

data+code: https://t.co/TGAzzUusOw

h/t @dhmontgomery (title+#rstats tips), @_cingraham (title+viz inspiration), @thejefflarson (emoji) pic.twitter.com/7PEkCT1eFP

— Christine Zhang (@christinezhang) July 3, 2018

#rstats tip of the day:

Do you often find yourself typing something like `paste0(round(number * 100, 1), "%")` to change your decimals to display as percents? Try `percent(number)` from the scales package instead! pic.twitter.com/S8nrUgszOD

— Emily Robinson (@robinson_es) July 3, 2018

Brush up on @thomasp85's animation 🔥 step-by-step before #useR2018
🎞 “Taking control of animations in R and demystifying them in the process”https://t.co/Qoeeqq27Oc #rstats pic.twitter.com/KDoJ7ILvB2

— Mara Averick (@dataandme) July 3, 2018

Tried my hand at creating an animated plot for #TidyTuesday this week. I've been needing to use a lot of #spatial libraries lately for work. Great way to practice. Also tried to be a little more analytical with a loess regression #rstats #tweenr #tidyverse https://t.co/FBTXN7JfVo pic.twitter.com/zhhmsGYvVk

— Dylan McDowell (@dylanjm_ds) July 3, 2018

Do you want to train Latent Dirichlet Allocation the same way as a Variational Autoencoder? TensorFlow Probability now includes an example on how to do this! https://t.co/GcynEfE7nA

— Michael Figurnov (@mfigurnov) July 2, 2018

I just published “Train on Google Colab and Run on the Browser: A Case Study” https://t.co/TSaC7Lnw1k

— Zaid Alyafeai (زيد اليافعي ) (@zaidalyafeai) July 2, 2018

That's an awesome ML ensemble Library! A user of mlxtend took the ensemble methods to the next level and added a computational graph approach to handle parallelization: "ML-Ensemble, a Python library for memory efficient parallelized ensemble learning" https://t.co/lc2SxApYWl

— Sebastian Raschka (@rasbt) July 3, 2018

💥 cool gallery of ggplot-related pkgs:
📈 "All Your Figure Are Belong to Us" 👾 @yutannihilat_enhttps://t.co/v8u5GfAsle #rstats #dataviz pic.twitter.com/aiz1Ul0h53

— Mara Averick (@dataandme) July 3, 2018

We’re hyped about hyperbolic entity embeddings - download our embeddings in #GloVe format from https://t.co/TZbNeJ7Nzi How can we put hyperbolic embeddings to work for you? Applications and feedback appreciated!

— Beliz Gunel (@belizgunel) June 28, 2018

One of my favorite little Python tidbits is that zip() is its own inverse...

>>> data = [(1, 2, 3), ('a', 'b', 'c')]
>>> zipped = zip(*data)
>>> unzipped = zip(*zipped)
>>> list(unzipped)
[(1, 2, 3), ('a', 'b', 'c')]

— Jake VanderPlas (@jakevdp) July 3, 2018

"Cambridge Analytica didn’t convince decent people to become racists; they convinced racists to become voters." Great article by @doctorow https://t.co/bdw0x3BLEE

— Andrew Barss (@andrewbarss) July 3, 2018

20th Century STEM Departments https://t.co/eH3XWgb455

— hardmaru (@hardmaru) July 3, 2018

"If you optimize everything, you will always be unhappy."
- Donald Knuth pic.twitter.com/VXqCO7lcje

— Francesc (@francesc) July 3, 2018

New blog post, co-written with @skyetetra! 12 red flags to watch out for in data science interviews 🚩https://t.co/hM2E7I46Da pic.twitter.com/jFVA7mmjjU

— Emily Robinson (@robinson_es) July 3, 2018

the problem with previous attempts at drag-and-drop machine learning interfaces is that they don't make the hard parts of ML easier. - @jeremyphowardhttps://t.co/JjmmrmhrO9 @jjvincent pic.twitter.com/deYg4uLLar

— Rachel Thomas (@math_rachel) July 4, 2018

I think the analysis misses a major confounder: most people use R primarily for exploratory graphics where (I would argue) the importance of design is less because the drawer of the plot is the reader of the plot. In EDA being able to rapidly iterate is of paramount importance

— Hadley Wickham (@hadleywickham) July 3, 2018

I had no idea that Stylish, the popular CSS userstyle browser extension, was collecting my complete browser history, including sites scraped from Google results. Instant uninstall. https://t.co/5X6hH5XXKY

— Andy Baio (@waxpancake) July 3, 2018

Google has been trying to make Gmail a platform for outside developers to build email apps and add-ons. In doing so, it’s given hundreds of companies the ability to read people’s email https://t.co/1Y4CroiidP

/1

— Doug MacMillan (@dmac1) July 2, 2018

Reading a paper on misconceptions about private browsing.
Oh dear. https://t.co/cjUNTpWk5H pic.twitter.com/cH3fGuSp4B

— Martin (@mshelton) July 3, 2018