Yet another Facebook “controversy” (to put it mildly) surface today. Udacity and Google launch free courses. And people debating what constitutes “Real Research” (check the first tweet in the Miscellaneous section).
[ethics] Facebook Sharing Data
Facebook confirms data sharing with Chinese companies https://t.co/7tnlS6aTPe
— Reuters Top News (@Reuters) June 6, 2018
After we published -- after a day of stonewalling -- Facebook finally confirmed this. But it has not answered key questions:
— Tony Romm (@TonyRomm) June 5, 2018
1. Have these integrations been wound down yet, or no?
2. What were the specific use cases?
3. While Huawei stored data on device, what about the others? https://t.co/H7itU0orby
[ethics] AI ethics and regulation
Interview on AI ethics and regulation, the dangers and distractions of AI hype cycles and how chat bots will become more and more useful for businesses as NLP continues to evolve.https://t.co/zH9FYVTTth pic.twitter.com/Q1N1Mggq2S
— Richard (@RichardSocher) June 6, 2018
MusicVAE
Take some time and enjoy the wonderful sounds of an ML model playing General MIDI arrangements for your listening pleasure: https://t.co/9bOlpjPZnn
— Ian Simon (@iansimon) June 5, 2018
Drone Surveillance System
AI-based real-time drone surveillance system "for Violent Individuals Identification". AI + cloud + drones = flying panopticon. Caveat: Cloud dependence for real-time identification.
— Jack Clark (@jackclarkSF) June 5, 2018
Vid: https://t.co/9V6ujEFWbh
Paper: https://t.co/0rYJK768L0
All-star panel @NAACLHLT
(Long thread. Click on the tweet to read the full thread.)
All-star panel at the generalization in deep learning workshop at @NAACLHLT #Deepgen2018 pic.twitter.com/gbxuDDubP0
— Sebastian Ruder (@seb_ruder) June 5, 2018
Standard CS degree does not have a single lecture on how to design an experiment. In grad school, vast % of PhD students have to do experiments w/o having ever been taught on how to do this before. Problems are unsurprising.
— Sebastian Ruder (@seb_ruder) June 5, 2018
On Blog Posts Promoting Papers
(specifically papers by big corporate labs) (Long thread. Click on the tweet to read the full thread.)
Excessive use of promotional blog posts promoting papers by big corporate labs to blast out (often) insignificant papers to millions of fan-{boys,girls}, who then preferentially cite this work in squillions of papers that make it to arXiv (if not peer review) should worry us.
— Zachary Lipton (@zacharylipton) June 5, 2018
Here's one example of a problematic blog post: https://t.co/motE7oHLj7 - I suspect part of why the negotiation bot story blew up when news outlets published hyperbolic articles on how FB had to shut down its chatbots after they went rogue (3/N)
— Zachary Lipton (@zacharylipton) June 5, 2018
World Cup Prediction
With the #FifaWorldCup2018 kicking off next week, I considered a number of variables to create the official @kdnuggets World Cup predictions https://t.co/FKo4JU2aIb pic.twitter.com/AWI7OqaFm9
— Dan Clark (@Dan_Clark5) June 5, 2018
Notable Research
Cyberattack Detection using Deep Generative Models with Variational Inference
— Jeremy Jordan (@jeremyjordan) June 5, 2018
https://t.co/Kj68urhjNB
detect anomalous behavior using reconstruction error of a variational autoencoder. model is capable of learning in an online setting although threshold values may need tuning. pic.twitter.com/0Mulps72El
"Playing Atari with Six Neurons," Cuccu et al.: https://t.co/o24EGI4AUS
— Miles Brundage (@Miles_Brundage) June 6, 2018
(6-18 for decision-making - additional computation needed for state representation which is separate but jointly trained)
"Multi-Modal Methods: Image Captioning" 🌟
— ML Review (@ml_review) June 5, 2018
By The M Tank (w/ @beduffy1)
A reflective narrative of key ideas and milestones in Image Captioning for the last 5 years: from Translation through Attention to Reinforcement Learning and beyond.https://t.co/3pvIrUAOE0 #ComputerVision pic.twitter.com/KnIuUbhK4A
Modern deep CNNs are unable to generalize over tiny image transformations such as translation, scale, contrary to widespread belief: https://t.co/xtc0Us0b06 @filippie509 @GaryMarcus
— Max Little (@MaxALittle) June 5, 2018
Tutorials and Resources
Just came across the markdown journal for the fast-deepnets project. Really nice to read through a journal taking through the project's evolution.
— Soumith Chintala (@soumithchintala) June 5, 2018
I usually maintain such a journal per project, but privately and not as pretty.https://t.co/Sa74L1yYEQ
gganimate has reached the gapminder state of its development as transition_time() has been added #rstats
— Thomas Lin Pedersen (@thomasp85) June 5, 2018
p + ... +
transition_time(year) +
ease_aes('linear') pic.twitter.com/Hz8azd1lio
Comprehensive Repository with 100’s of #DataScience and #MachineLearning Resources for #DataScientists: https://t.co/qz1nRPOxJj #abdsc #BigData #AI #DeepLearning #NeuralNetworks #Statistics pic.twitter.com/G8bA0oPutN
— Kirk Borne (@KirkDBorne) June 5, 2018
Udacity and Google launch free career courses for interview prep, resume writing and more https://t.co/9M9uz2vHMl
— TechCrunch (@TechCrunch) June 5, 2018
Announcing Public Git Archive (https://t.co/880kDCFhjt) for static analysis and #MLonCode
— source{d} (@sourcedtech) June 5, 2018
3 Terabytes of GitHub repositories (all projects >50 stars) with meta data, stored as rooted repositories, and easily downloadable over HTTP with the pga tool (https://t.co/yDxXyqvnEZ). pic.twitter.com/MaBNq4fIeC
New sphinx-gallery v0.2 is out! Automatically create beautiful galleries from your python examples. Now with auto-generating Binder links! https://t.co/pIRIq9WOg2 pic.twitter.com/flouXfldPW
— Chris Holdgraf (@choldgraf) June 5, 2018
woah Jetbrains launched a web based data analysis ide!https://t.co/bLhhrvfJJ6
— Prashant Deva (@pdeva) June 5, 2018
Super excited to announce MLflow, a new open source Machine Learning platform from Databricks to manage the complete machine learning lifecycle: https://t.co/nBDu5KNqwv
— Matei Zaharia (@matei_zaharia) June 5, 2018
Python
Meaning to learn Python or brush up on basics? Join our 7-day Python Challenge, starting June 11th. Ssssign up here: https://t.co/IlaL5mIyHH 🐍🐍🐍
— Kaggle (@kaggle) June 5, 2018
Itertools in Python 3, By Example via @realpython https://t.co/UwQDkGwUkt #python pic.twitter.com/6q2YtYD8dX
— Python Weekly (@PythonWeekly) June 5, 2018
How to debug Python, for beginners:
— Denny Britz (@dennybritz) June 5, 2018
1. Carefully read the exception message. Repeat it to yourself 3 times, slowly.
2. Step away from the computer. Ideally, lie down somewhere.
3. Close your eyes, take 10 deep breaths.
4. You will have found your mistake.
Miscellaneous
(Long discussion thread. Click on the tweet to read the full thread.)
I'm surprised there are still so many academics that think that Real Research must be in the form of papers and equations, not software.
— Jeremy Howard (@jeremyphoward) June 5, 2018
So many brilliant people driven out of academia by this, which reinforces the cycle. https://t.co/jluKxiflP0
Great long read in @theipaper by @robhastings, looking at how Streams, our secure mobile app, is working at @RoyalFreeNHS. Well worth a read to understand the enormous potential for positive impact of digitisation in healthcare! :)https://t.co/aLnvxZOjHF
— Mustafa Suleyman (@mustafasuleymn) June 5, 2018
Should be noted the key role of #XGboost (@tqchenml !) and have #BDT tackle directly the 4-momenta (instead of handcrafted variables). Both emerged during our #higgsml #kaggle challenge. Took 4 years, because, eh!, everything takes longer #irl but it was worth it. #hepml https://t.co/tswx4K5L7S
— David Rousseau (@dhpmrou) June 4, 2018
"From AI to ML to AI: On Swirling Nomenclature & Slurried Thought" - a new installment related to "the AI Misinformation Epidemic" - on @approxcorrect https://t.co/4UrIfJubbI
— Zachary Lipton (@zacharylipton) June 5, 2018
"What do you do?"
— karl rohe (@karlrohe) June 5, 2018
"I'm a Statistician."
"Oh wow, I hated that class."
"Sorry. We are bad at teaching."
"I bet you've seen some crazy statistics!"
“uhhhhh.” https://t.co/JztjgO0TKP
For the next ±3 months you can send in your (own, or another's) favorite #dataviz made in the past year! 🎖️ And please do! I get loads of inspiration from browsing through the eventual Showcase that comes out 😃 https://t.co/j4avJQuJk5
— Nadieh Bremer (@NadiehBremer) June 5, 2018