Quite a few tweets popped out this day with #NAACL2018 hashtag. I’m not organizing them into one single section, though. They are scattered in different sections.
The Roads to Rome
The roads to #Rome - U.S. version. These roads lead to the nearest city named Rome. #dataviz
— Randy Olson (@randal_olson) May 31, 2018
More maps at https://t.co/xeINtIvfWg pic.twitter.com/cHIYwdTKko
“truly-ism”
"Deep learning has succeeded primarily by showing that certain questions or tasks we thought were difficult are in fact not. It has not addressed the truly difficult questions that continue to prevent us from achieving human like AI" - Judea Pearl - The Book of Why
— Gigasquid (@gigasquid) June 3, 2018
This is a classic case of "truly-ism". It turns out every problem we solve was solvable. One day, someone will answer the core questions of intelligence and someone will say, "we thought it was difficult but it was in fact not." https://t.co/7OrXHZJkRf
— Thomas G. Dietterich (@tdietterich) June 3, 2018
My hypothesis is that the function of consciousness is self-monitoring and forward projection for safety and robustness. We are creating AI systems with these functions out of engineering necessity. We will look back and say "it was easy".
— Thomas G. Dietterich (@tdietterich) June 3, 2018
[Ethics] On Gender Imbalance (Cont.)
Continuing the topic from Yesterday
There are no excuses for any #rstats conference to have all-male speakers & committee. It is of gravest concern to our #rstats community. #rfinance2018 has fallen short. We are working with 2 of their sponsors, @RConsortium and @rstudio, to ensure better practices going forward. https://t.co/kyBTQP3DwL
— R-Ladies Global (@RLadiesGlobal) June 2, 2018
Me pointing the conference organizers to all the great women in data science who they could have invited like pic.twitter.com/KjEAM0AGKo
— Cynthia Lee (@cynthiablee) June 2, 2018
Diversity call to action! We are reacting and being proactive in moving forward together. Come and join us. #rfinance2018 #rladies @RLadiesChicago @RLadiesGlobal #wimlds pic.twitter.com/t0LiBvuBsz
— Amy Yang (@ayanalytics) June 2, 2018
We're actively working with the #rfinance2018 organizers to make their conference more inclusive toward all! This is a first step, and there's lots more to come. (That's the power of having #RLadies on your team.) #rstats https://t.co/PAH5Cfy2My
— Angela Li (@CivicAngela) June 2, 2018
Just going to put this twitter list of over 1400 women in data science and related fields right here 😉https://t.co/hTEsK2R2sX
— Data Science Renee (@BecomingDataSci) June 2, 2018
(Let me know if you want to be added!)
Being underrepresented is about numbers; being marginalized is about treatment. White women are underrepresented but usually not marginalized, which is why white women != diversity.
— Alaina @ #JSConfEU 🇩🇪 (@alainakafkes) June 2, 2018
Props to @KimCrayton1 for making it crystal clear! 👏 #JSConfEU
[Ethics] Facebook Kills Trending Topics
In big shift: @Facebook kills trending topics to address "difficulty of relying on computers, even artificial intelligence, to make sense of the messy human world without committing obvious, sometimes embarrassing and occasionally disastrous errors." https://t.co/rqZTFK4C80
— Eileen Donahoe (@EileenDonahoe) June 1, 2018
So, wait. AI makes “embarrassing and occasionally disastrous” errors when we ask it to surface information, but it’s OK for suppressing information? Isn’t the difference just that the first one has unavoidable public transparency, so we all know about the errors? https://t.co/Ao430a0G8W
— Daphne Keller (@daphnehk) June 1, 2018
Using filters to find and delete “terrorist” content. The thing Zuckerberg kept telling Congress they would do.
— Daphne Keller (@daphnehk) June 2, 2018
Notable Research
Learned a lot about LSTM behavior -- in very different ways -- from two excellent @acl2018 papers: Sharp Nearby, Fuzzy Far Away... by @ukhndlwl, He He, Peng Qi, and @jurafsky, and LSTM as Dynamically Computed... by @omerlevy_ , @kentonctlee, @nfitz, @lukezettlemoyer.
— John Hewitt (@johnhewtt) June 2, 2018
Really cool work from the Brunskill lab at Stanford on using model-based RL! https://t.co/s7bbugQq01
— Karan Goel (@krandiash) June 1, 2018
Distributional regression forests on arXiv https://t.co/JWmbr4n2jh #rstats #DataScience
— R-bloggers (@Rbloggers) June 2, 2018
Fast Abstractive Summarization with Reinforce-Selected Sentence Rewriting #acl2018
— ML Review (@ml_review) June 2, 2018
By @YenChunChen4
Fast summarization model that selects salient sentences and then rewrites them abstractively (i.e., compresses and paraphrases)https://t.co/kpX41Wbqgq pic.twitter.com/ZokX15wCui
Peyman Passban, Qun Liu, Andy Way (presented by @seb_ruder ): Improving Character-based Decoding Using Target-Side Morphological Information for Neural Machine Translation https://t.co/Cx2PuxNgSd #NAACL2018
— Yuval Pinter is livetweeting from #NAACL2018 (@yuvalpi) June 2, 2018
Tutorials and Resources
Slides for my #cascadiarconf talk on contributing to R packages are here: https://t.co/z4QvUiNOTg
— Kara Woo (@kara_woo) June 2, 2018
✍🏼advice from @kara_woo for contributing to #rstats packages: improve documentation. Did this for the first time last week to give more dplyr::na_if examples because I could never use it on data frames that included dates 😖 #cascadiarconf https://t.co/xsVjGgwobj pic.twitter.com/fKRo4RC0fT
— Alison Hill (@apreshill) June 2, 2018
Great way to discover cool #rstats projects: search @dataandme's timeline for code-throughs! - @apreshill's keynote at #cascadiafconf pic.twitter.com/CkAgUSHNcN
— Caitlin Hudon👩🏼💻 (@beeonaposy) June 2, 2018
Free @utknightcenter MOOC "Data Visualization for Storytelling and Discovery", taught by @albertocairo. You'll learn how to use graphs, maps, charts, diagrams to extract meaning from large amounts of data & how to use #dataviz to tell stories https://t.co/bqX0E1k5PF pic.twitter.com/wHbiNIwDJQ
— GIJN (@gijn) June 2, 2018
SpaceNet is an online repository of freely available satellite imagery, co-registered map data to train algorithms, and a series of public challenges designed to accelerate innovation in machine learning using geospatial data. https://t.co/asyQfiwXrC
— Christopher (@communicating) May 31, 2018
Looking for #naacl2018 papers? Avoid the browser. David Vilar and I wrote a tool for querying and managing official bibtex entries and their associated PDFs. Also supports arXiv searching. "bibsearch man" for more info. pic.twitter.com/VX2F2zrYeV
— Matt Post (@mjpost) June 2, 2018
Understanding NBA Foul Calls with Python by Austin Rochford - https://t.co/OzX5keU32h. This talk is a case study in using open source Python packages to analyze these reports in order to understand the relationship between game dynamics, player abilities, and foul calls.
— Python Software (@ThePSF) June 2, 2018
Deep Probabilistic Methods with PyTorch - Chris Ormandy https://t.co/lORbgRCD74 #pytorch #deeplearning #machinelearning #ml #ai
— PyTorch Best Practices (@PyTorchPractice) June 2, 2018
Very interesting tutorial on Socially Responsible NLP at #NAACL2018 by Yulia Tsvetkov, Rob Voigt and Vinodkumar Prabhakaran. Great conversations and some very interesting work presented. pic.twitter.com/4UgiS4Sp07
— Kartikay Khandelwal (@kakemeister) June 1, 2018
keras-applications - Reference implementations of popular deep learning models. https://t.co/q8DRCtO5nk
— Python Trending (@pythontrending) June 2, 2018
The reason why I attempt to recreate the plots given every week, is that it's my goal to attack the data and make the plot as quick as I can. I have gotten much better at my data tidying/visualization skills this way. #TidyTuesday #rstats #tidyverse #comic https://t.co/m7qorlnoey pic.twitter.com/EMOa5MRvs5
— Dylan McDowell (@dylanjm_ds) May 30, 2018
Miscellaneous
— Amanda Stent (@astent) June 1, 2018
@mrogati's medium post https://t.co/ThrzSEa1Oa has become my No.1 post to share when talking to cross-functional partners or aspiring DS. Trying to imprint the triangle on people's mind
— Robert Chang (@_rchang) June 3, 2018
“One reason Japanese is a difficult language to [machine] translate, is that people often omit subjects because Japanese people understand each other without mentioning subjects.” https://t.co/rr1WxXCMRc
— hardmaru (@hardmaru) June 3, 2018
Country-by-country life expectancy since 1543. #health #datavizhttps://t.co/8JI9UKDnWl pic.twitter.com/FvqWlpjJo4
— Randy Olson (@randal_olson) June 2, 2018
Design is code. Donald Knuth's gift: complete integration: type image math,
— Edward Tufte (@EdwardTufte) December 31, 2016
whatever it takes. ACP4,pre-fascicle 5C,Dec10,2016 #dataviz #ddj pic.twitter.com/VCDRZKnWfU
“An algorithm must be seen to be believed.” — Donald Knuth
— hardmaru (@hardmaru) June 3, 2018
Humanitarians urgently need professional #standards & #ethics for #HumTech #HumData #CommIsAid activities. We @HHI_Signal researched + wrote the Signal Code: Ethical Obligations for #Humanitarian #Information Activities in response to this need: https://t.co/GLz9dPvHFr @eu_echo pic.twitter.com/Xxlm6epA9X
— Stuart Campo (@stucampo) June 1, 2018
LaTeX is so much better when you have emoji. pic.twitter.com/SYxgUXpWYN
— Mike Bostock (@mbostock) June 2, 2018
“Instead of trying to solve every problem with a single language, in my case: C++, and become an entrenched ‘C++ expert’, it is much more enjoyable and productive to learn a few different languages and pick a language that naturally fits a problem.” https://t.co/yoHST4RVRa
— hardmaru (@hardmaru) June 3, 2018
Want a good-ol'-fashioned hard copy of the NEWSROOM summarization dataset (https://t.co/sKrd34Z8x8)? Find Max Grusky at #NAACL2018 and get 1.3M article-summary pairs on a bespoke flash drive, limited supplies! -- TALK ON SUNDAY, 11:06 in Empire B #NLProc https://t.co/tqYKCZDjsg pic.twitter.com/cGy5NDOdg8
— Yoav Artzi (@yoavartzi) June 2, 2018
Don’t put too much pressure on yourself when learning! I strongly agree with @apreshill that learning projects should be fun and don’t always need to be Super Serious. #cascadiarconf pic.twitter.com/BxWBgPOIL2
— Kara Woo (@kara_woo) June 2, 2018
Having a great time at @cascadiarconf ! I was inspired to make this short, mostly inaccurate summary of the deep learning workshop from @DynamicWebPaige #rstats #cascadiarconf pic.twitter.com/xNuKKval51
— Jay Lee (@jaylee_tx) June 2, 2018
Neural-net Generated Joke
Here's a neural net-generated joke:
— Janelle Shane (@JanelleCShane) June 2, 2018
Why did the chicken cross the road? To screw in a light bulb.
for this one I used textgenrnn. It's basically a char-rnn.
— Janelle Shane (@JanelleCShane) June 3, 2018