alextp

7 ms·

1.
by alextp 2mo ago
When claire stopped writing the tgif emails I was very sad, among others I talked to. When I heard, years later, about all this bs she went through, it helped crack the illusion for me. So sad.
2.
by alextp 4mo ago
The flip side is redistributive pensions require an ever growing population and most European pension systems will go bankrupt within a couple of decades given current birth and immigration rates.
3.
by alextp 4y ago
Choose the top N according to the proxy objective and then use the real objective to choose the best out of those N candidates.
4.
by alextp 7y ago
Claire deWitt and the city of the dead, by Sara Gran is a good example. Very colorful language. Also the first volumes of the Berlin Noir "trilogy".
5.
by alextp 7y ago
How is brain separate from the rest of google?
6.
by alextp 8y ago
The historic picture makes a little more sense (though this is not something a 5yo would understand). We call these things embeddings because you start with a very high dimensional space (image a space with one dimension per word type, wher
7.
by alextp 8y ago
I've contributed to autograph and would love to answer any questions.
8.
by alextp 9y ago
Cool! Did you try using tensorflow's eager execution?
9.
by alextp 9y ago
I think Keras is a real deal framework. It provides a higher-level API than most other frameworks, but it has pretty sweet portability of models across frameworks and platforms and most research papers are implementable in Keras without too
10.
by alextp 9y ago
We're still fairly early in the project, so for now threading is the only supported way. We can do better, however, and we're working on ways to leverage the hardware better (for example, if you have no data-dependent choices in y
11.
by alextp 9y ago
Ah, I didn't know SavedModel didn't work in android. I think freezing is still the way to go there? I'm sorry, I don't personally work on the mobile side of things.
12.
by alextp 9y ago
Did you try using SavedModel? It should be seamless to use downstream with tensorflow serving and it's not that hard to get estimators to spit those out.
13.
by alextp 9y ago
You can read out more about it in the blog post ( https://research.googleblog.com/2017/10/eager-execution-impe... ) or the README ( https://github.com/tensorflow/tensorflow/tree/mast
14.
Eager Execution: An imperative, define-by-run interface to TensorFlow(research.googleblog.com)
125 points by alextp 9y ago|35 comments
15.
by alextp 12y ago
I write tests not to convince myself my code is correct (it often is, and repl and ad-hoc testing are more than enough to make sure it does the right thing now) but to prevent myself & future others from breaking it as they maintain it
16.
by alextp 13y ago
The averaged perceptron is the one to which the proof applies :-)
17.
by alextp 13y ago
Regarding the perceptron, most modern texts also have a regret analysis for the perceptron, which coupled with an online-to-batch conversion tells you how well do you expect a perceptron to perform on unseen data after a single pass, and it
18.
by alextp 13y ago
ML PhD student here. The reason why this is different is that the parallel monte carlo simulations are running on different subsets of the data in each machine, and then averaged. It is not obvious that this can work at all in some cases. T
19.
by alextp 14y ago
Honestly, more people should use stabilizer http://plasma.cs.umass.edu/emery/stabilizer . The best can still be an outlier.
20.
by alextp 14y ago
Fun fact: this is pretty much the same dynamic program as CKY, the parsing algorithm that can parse any context-free language in the appropriate normal form (only productions looking like A->BC or A->a) in time cubic on the length o
21.
New backend optimizer for scala(magarciaepfl.github.com)
1 points by alextp 14y ago|0 comments
22.
by alextp 14y ago
The fallacy behind arguments like this is that when someone says that we're "proposing to push people to speak at conferences who aren't necessarily the most knowledgeable speakers" it is implicitly assumed that it is known how to rank the
23.
by alextp 14y ago
A foot a second is off by several orders of magnitude. It's closer to 300 thousand km per second in vacuum, and around 200 thousand km per second in fiber.
24.
by alextp 14y ago
The above construction is a not a breakthrough and it is how I was taught in my approximate algorithms class. Also the way the optimization problem is usually defined is not as "give a circuit with minimal cost" but "what is the cost of the
25.
by alextp 14y ago
A temp agency?
26.
Sci vs. Sci: Evidence of Facebook Voting Fraud in Bespoke-Suit Social Promotion(joseph.turian.com)
7 points by alextp 14y ago|0 comments
27.
by alextp 14y ago
Regarding finding the higgs boson, regardless of whether you agree that it's what they're actually doing, the main people actually building the first quantum computer are D-Wave, a private company.
28.
by alextp 14y ago
It's not about splitting the task into smaller pieces. It's about factoring out the parts of the task that need synchronization among all machines into one specific subroutine (groupBy) which makes mapreduce so powerful. If you speak with p
29.
by alextp 14y ago
Mapreduce as a concept goes beyond lisp implementations. On the surface it might seem like the point of mapreduce is expressing computations in terms of map and reduce functions. It isn't. The point of mapreduce is reducing the problem of h
30.
by alextp 14y ago
But using personally identifiable data for targeting (which facebook seems to be doing) is entirely different from doing personally identifiable targetting (which no one is doind). If all the information that leaves facebook servers is whic

More ›