dandermotj

I'm currently doing Andrew Ng's course and wrote a short review of course 1 here [1]. If you have the mathematics and statistics background and want to go through that rigorously enough then it's exceptional. If you are not the mathematically inclined it is still accessible but I can imagine fast.ai being more appropriate, although I have no personal experience.

[1] https://dandermotj.github.io/post/review-deeplearning-ai-cou...

In the EU we have the incoming GDPR to legislate for (and penalise) data breaches like this. This directive is very clear and detailed on how data should be collected, securely stored and disposed of. US law is a decade behind the EU.

Googling LeNet, Reset, AlexNet or other well known architectures will provide you with links to the network weights.

I'm always a bit disappointed that yhat indiscriminately used 'ggplot' for the python package name. Using a variation on the name would have been more considerate.

This is definitely a scientific paper. Pretty much no scientific paper comes with source code and the majority of scientific papers are not reproducible without an entire university department of resources anyway.

As an Irish and European person, I want you to know that your comment is misinformed and misleading. Global tax avoidance and evasion by multinationals has no relation to the formation of the EU, and the Irish government are at a gigantic economic loss (not advantage) as a result.

Breaking the spirit of the law, but not the letter.

I like to think about it this way. By not explicitly imposing a prior, you are implicitly imposing a prior that each item will receive no votes. This is totally non sensical because of course these items will get votes.

Just because we don't know what the true value of p will be doesn't mean we don't have some expectation. If I asked you what you expect the popularity of a given item will be, you won't say 0, you'll say something like the average. So why assume all items will have 0 votes in our model?

Say hello to the military-industrial complex

Consumption through spending is economically beneficial to everyone. You misunderstand what 'trickle down economics' is.

Clickbait, but great results all the same

Thanks :D

Hi Chris, firstly thanks for all the work you've done publishing brilliant articles on supervised and unsupervised methods and visualisation on your old blog and now in Distill.

This question isn't about feature visualisation, but I though I'd take the chance to ask you, what do you think of Hinton's latest paper and his move away from neural network architectures?

Sparse heterogeneous data is often the type of data stored in NoSQL dbs. Modelled in a relational way, this produces many tables with many NULL fields, while keeping it in a key value format is neat and tidy.

I'd recommend the paper What Goes Around Comes Around[1], the first paper in Readings in Database Systems[2]

[1] https://scholar.google.com/scholar?cluster=73661829057771494... [2]redbook.io

And obesity levels are higher than ever too. Dismissing this article's message because we are living longer is missing the point. We are living longer despite our poor physical health.

That's a P/E of 60... Also this:

> In exchange, Patreon takes only a tiny 5% cut.

I hope this is sarcasm, because if I had to pay 5% to make any other transaction I'd be fuming. With a marginal cost of facilitating a new patron near zero, I don't think creators or patrons will suffer this as Patreon grows.

Which is more legible and understandable?

    f(a, g(h(x), b))
    h(x) %>% g(b) %>% f(a, .)

"Thus, programs must be written for people to read, and only incidentally for machines to execute."

You are loosely describing collaborative filtering [1], a very common recommendation technique.

[1] https://en.wikipedia.org/wiki/Collaborative_filtering

This is a fantastic book, I loved it.

I'd tend to agree. Any large corporation or financial institution that has adopted Hadoop has done so for very good reasons.

If anyone wants to know how deep down the rabbit hole this stuff goes, they should read this blog post [1] on writing a Levenshtein Automaton to speed up fuzzy matching in Lucene by 100 times. It gets deep!

[1]: http://blog.mikemccandless.com/2011/03/lucenes-fuzzyquery-is...

Always impressed with the expressiveness and flexibility of the tidytext framework for analysing text. Well done Julia!

Julia and David Robinson's book Tidy Text Mining with R[1] has plenty more great text mining examples like this one.

[1] http://tidytextmining.com/

Does anyone have a place/site for browsing through new/popular utilities? I've been on Ubuntu for the last year and I love it, but reading this I just realised I don't have anywhere to browse or find things like this.

It's very easy to set up an RStudio server. I used to run one on Azure. Took minutes to set up and you get the whole RStudio suite, including R notebooks.

Probably the one tool missing from this list is Impala, which is essentially Hive's successor. Uses the same metastore and runs an order of magnitude faster. Almost the same flavor of SQL too.

Please god let this be taken up! I'm so fed up of

     sudo service network restart

I'm definitely not a designer, but I will say I love the Ubuntu fonts

From the title I thought this was a machine learning paper... ¯\_(ツ)_/¯

The latter

I studied statistics - my point was that statistics is taught in a linear manner, starting with distributions and hypothesis testing (p-values) and then move onto more advanced treatments like Bayesian stats.