Hacker News: alextp

New comment by alextp in "The Voice of Google"

alextp — Mon, 20 Jul 2026 16:03:28 +0000

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.

New comment by alextp in "The death of the brick and mortar toy store"

alextp — Fri, 22 May 2026 12:44:25 +0000

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.

New comment by alextp in "Measuring Goodhart’s Law"

alextp — Wed, 13 Apr 2022 22:01:50 +0000

Choose the top N according to the proxy objective and then use the real objective to choose the best out of those N candidates.

New comment by alextp in "Great Noir Lives and Dies On Dialogue"

alextp — Tue, 10 Sep 2019 23:39:34 +0000

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".

New comment by alextp in "Why Google+ Failed"

alextp — Thu, 20 Jun 2019 18:07:56 +0000

How is brain separate from the rest of google?

New comment by alextp in "Wittgenstein’s theories are the basis of all modern NLP"

alextp — Wed, 09 Jan 2019 22:58:31 +0000

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, where each word is a unit vector in the appropriate dimension) and then approximate distances between sentences / documents / n-grams in this space using a space with much smaller dimensionality. So we "embed" the high dimensional space in a manifold in the lower dimensional space.

It turns out though that these low dimensional representations satisfy all sorts of properties that we like which is why embeddings are so popular.

New comment by alextp in "AutoGraph converts Python into TensorFlow graphs"

alextp — Tue, 17 Jul 2018 22:31:51 +0000

I've contributed to autograph and would love to answer any questions.

New comment by alextp in "Generative Adversarial Networks Code in PyTorch and Tensorflow"

alextp — Fri, 05 Jan 2018 22:23:10 +0000

Cool! Did you try using tensorflow's eager execution?

New comment by alextp in "Eager Execution: An imperative, define-by-run interface to TensorFlow"

alextp — Wed, 01 Nov 2017 02:22:59 +0000

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 much trouble.

New comment by alextp in "Eager Execution: An imperative, define-by-run interface to TensorFlow"

alextp — Wed, 01 Nov 2017 02:21:36 +0000

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 your model we can enqueue kernels in parallel on all GPUs in your machine at once from a single python thread, which will perform much better than explicit python multithreading).

Stay on the lookout as we release new experimental APIs to leverage multiple GPUs and multiple machines.

New comment by alextp in "Eager Execution: An imperative, define-by-run interface to TensorFlow"

alextp — Tue, 31 Oct 2017 19:32:52 +0000

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.

New comment by alextp in "Eager Execution: An imperative, define-by-run interface to TensorFlow"

alextp — Tue, 31 Oct 2017 19:10:55 +0000

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.

New comment by alextp in "Eager Execution: An imperative, define-by-run interface to TensorFlow"

alextp — Tue, 31 Oct 2017 18:13:58 +0000

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/master/tensorf... ). This is still a preview release, so you may hit some rough edges.

Looking forward to your feedback as you try it out.

Eager Execution: An imperative, define-by-run interface to TensorFlow

alextp — Tue, 31 Oct 2017 17:55:36 +0000

Article URL: https://research.googleblog.com/2017/10/eager-execution-imperative-define-by.html

Comments URL: https://news.ycombinator.com/item?id=15595123

Points: 125

# Comments: 35

New comment by alextp in "Is TDD Dead?"

alextp — Fri, 27 Jun 2014 15:21:14 +0000

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 in the future.

New comment by alextp in "Machine learning is easier than it looks"

alextp — Sat, 23 Nov 2013 15:26:52 +0000

The averaged perceptron is the one to which the proof applies :-)

New comment by alextp in "Machine learning is easier than it looks"

alextp — Thu, 21 Nov 2013 14:29:09 +0000

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's usually a very good estimate (the answer is on average about as well as it did on the examples in the training data).

New comment by alextp in "Bayes and Big Data: The Consensus Monte Carlo Algorithm"

alextp — Mon, 18 Nov 2013 19:06:41 +0000

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. Think, for example, a clustering model. If there are two clusters, but one machine calls them A B and the other machine calls them B A, averaging will give you useless results.

So the contribution of this paper is finding a set of models on which naive averaging works, and showing an efficient mapreduce implementation of it.

That said, I don't find the paper particularly interesting.

New comment by alextp in "GCC and C vs C++ Speed, Measured"

alextp — Thu, 21 Mar 2013 12:44:51 +0000

Honestly, more people should use stabilizer http://plasma.cs.umass.edu/emery/stabilizer . The best can still be an outlier.

New comment by alextp in "Dynamic Programming: Chain Matrix Multiplication"

alextp — Thu, 31 Jan 2013 15:37:27 +0000

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 of the sentence.