DataQuo.net (beta)new | comments | leaders | submitlogin
1.
Use LSTMs on time series [awesome TensorFlow tutorial] (github.com)
3 points by GChe 172 days ago | discuss
2.
In case you did not realize it, Nobody enjoys writing and maintaining data pipelines or ETL (stitchfix.com)
4 points by georges 365 days ago | discuss
3.
Much of the "data science" happening in industry is happening in isolation on data scientists' laptops (datadan.io)
3 points by kalil 355 days ago | discuss
4.
At the risk of disappointing you, know that ML does not solve all the problems. However, it allows us to provide answers to some of them.
2 points by Slim 167 days ago | discuss
5.
If you can write a for loop you can do statistics (speakerdeck.com)
3 points by georges 444 days ago | discuss
6.
Machine learning is not equivalent to R scripts (medium.com)
3 points by kalil 444 days ago | discuss
7.
Datasets — not algorithms — might be the key limiting factor to development of human-level AI (edge.org)
3 points by georges 447 days ago | discuss
8.
Tomorrow's developers will use machine learning like today's developers use sort() (twitter.com)
2 points by georges 351 days ago | discuss
9.
There is no process in data science like Agile is to software engineering or Lean is to manufacturing (dominodatalab.com)
2 points by kalil 365 days ago | discuss
10.
Today’s ‘Artificial Intelligence’ is tomorrow’s 'Data Science’ (beardedbrownman.com)
3 points by Kuite 508 days ago | discuss
11.
The more time spent in meetings, the more a data scientist (/analyst/engineer) earns (linkedin.com)
3 points by Kuite 520 days ago | discuss
12.
POCs are super easy. It's time to go to production (twitter.com)
3 points by Hyme 540 days ago | discuss
13.
Unfortunately, the predictive models that are most powerful are usually the least interpretable (machinelearningmastery.com)
6 points by vien 946 days ago | discuss
14.
The cost of a top, world-class deep learning expert is about the same as a top NFL quarterback prospect (gigaom.com)
5 points by vien 946 days ago | discuss
15.
Data engineers should often be hired before data scientists at startups (dataquest.io)
2 points by Kuite 520 days ago | discuss
16.
The problem is not ‘which model do we choose’ but ‘what action do we take’ (academia.edu)
5 points by test 964 days ago | discuss
17.
Hardware may actually become an accessory to data (economistgroup.com)
2 points by oula 540 days ago | discuss
18.
It's what you do with Data that makes it "Big" (aptitudesoftware.com)
3 points by Jebone 747 days ago | discuss
19.
Data Science is mostly an Engineering problem (youtube.com)
3 points by craigh 760 days ago | discuss
20.
Complex models are to Data Science, what “haute couture” is to the clothing industry : they are not made to be daily used, but are necessary (medium.com)
4 points by fire 918 days ago | discuss
21.
Whether or not you increase complexity for additional accuracy is not a data science decision. It’s a business decision (john-foreman.com)
4 points by juoilr 943 days ago | discuss
22.
Predictive modeling feeds back into systematic discrimination (john-foreman.com)
3 points by craigh 795 days ago | discuss
23.
Machine learning is a meritocracy (machinelearningmastery.com)
3 points by Jebone 846 days ago | discuss
24.
"Big Data isn't actually interesting without Machine Learning". Steve Jurvetson (slideshare.net)
3 points by craigh 847 days ago | discuss
25.
Each Kaggle challenge is like an odyssey … (medium.com)
3 points by craigh 907 days ago | discuss
26.
On deeplearning: "One does not simply learn DEEPLY" (twitter.com)
3 points by jilo 936 days ago | discuss
27.
Data science is a simple game in which you throw data at a bunch of methods and in the end random forests always win (twitter.com)
3 points by Clay 944 days ago | discuss
28.
The ultimate insult in Data Science : "it explains everything but predicts nothing" (quantumdiaries.org)
3 points by habit 946 days ago | discuss
29.
A prediction is always a missed target (ipython.org)
2 points by jilo 789 days ago | discuss
30.
Frankly, people are more interested in ImageNet than in CIFAR-10 nowadays (kaggle.com)
2 points by fire 822 days ago | discuss
More