transcranial

9 ms·

31.
by transcranial 9y ago
Totally, I think people have this weird sense of entitlement when it comes to high-quality datasets without the commensurate respect for how they're created or the level of effort that goes into them. Fei-Fei Li gives a good sense for
32.
by transcranial 9y ago
But it's already diverged enough IMO to warrant enjoying them separately. Might as well get the full experience... Plus without the source material to work from I'm skeptical as to just how great the TV ending will be. But I guess
33.
by transcranial 9y ago
A tremendous amount of trust is certainly involved. The brain itself doesn't actually have any pain receptors!
34.
by transcranial 9y ago
This is not at all uncommon: https://www.washingtonpost.com/news/speaking-of-science/wp/2... https://www.bustle.com/articles/35791-watch-inspiring-violin... I've watched a few mysel
35.
Uber CEO Kalanick's Resignation Is Not the Best Answer(forbes.com)
2 points by transcranial 9y ago|0 comments
36.
by transcranial 9y ago
Great video on the power of play as a driver of innovation/invention: https://ww2.kqed.org/mindshift/2017/04/13/how-play-is-at-the...
37.
by transcranial 10y ago
I just started reading the Nexus trilogy by Ramez Naam ( http://rameznaam.com/nexus/ ). It's quite good so far, and explores some of these concepts from a sci-fi bent.
38.
A.I. versus M.D.(newyorker.com)
3 points by transcranial 10y ago|0 comments
39.
by transcranial 10y ago
This is really exciting! Chris et al: have you guys seen Keras.js ( https://github.com/transcranial/keras-js )? It could probably be useful for certain interactive visualizations or papers.
40.
by transcranial 10y ago
There are definitely measurable population-level effects, like increased traffic accidents [1] and heart attacks [2]. And given that its purported benefits are completely unsubstantiated, it's amazing that DST still exists. [1] https:
41.
Kaggle Joins Google Cloud(blog.kaggle.com)
67 points by transcranial 10y ago|7 comments
42.
by transcranial 10y ago
IBM bought Merge Healthcare for $1B, presumably to gain access to vast amounts of their imaging data. Would love to see some developments on the deep learning front from them, but sadly I doubt it'll ever break through the fog of marke
43.
by transcranial 10y ago
Looks really amazing! Reproducibility of models and experiments is huge. It should be almost a requirement if one is going to publish results claiming SotA, etc. Seems like you could become the GitHub for deep learning in addition to the He
44.
by transcranial 10y ago
I've been using Stylish ( https://userstyles.org/styles/browse/github ) for this -- works pretty well.
45.
by transcranial 10y ago
Unfortunately the paper is in Nature, paywalled, instead of Arxiv, and data/code/model/weights inaccessible. While publishing in Science/Nature/NEJM/JAMA is definitely the right approach for deep learning to ga
46.
by transcranial 10y ago
I think medicine will over time morph from a field generally perceived as an intellectual one, to a largely humanistic one, like nursing. Most people, especially the HN crowd, vastly underestimate the importance of the human touch. A doctor
47.
by transcranial 10y ago
I actually played the amazing game Monument Valley after seeing it in House of Cards. Although from what I can recall this wasn't paid placement so much as the writers just loved the game.
48.
by transcranial 10y ago
They _do_, but they _shouldn't_. It's an inefficient allocation of skills and resources.
49.
by transcranial 10y ago
Have you seen Keras.js? It uses WebGL to run certain computations on the GPU. https://github.com/transcranial/keras-js
50.
by transcranial 10y ago
Yeah, setting aside whether his claim is actually true or not, why would this even be newsworthy or matter, if it were not for this context?
51.
Deep learning for predicting bone age on pediatric x-rays(demos.md.ai)
1 points by transcranial 10y ago|0 comments
52.
by transcranial 10y ago
Thanks. Nothing fancy with the network architecture diagrams. The layers are just div elements with the layer name as the id. There's a definition of inbound/outbound connections by layer names, extracted from the keras json confi
53.
by transcranial 10y ago
You mean with tensorflow or theano as the backend? They have all kinds of optimizations that isn't possible to replicate here yet. There is certainly room for optimization! Also, 1000-layer resnets should theoretically be possible, but
54.
by transcranial 10y ago
Thanks! For WebGL, credit goes to https://github.com/waylonflinn/weblas . I only really use GEMM, but it works quite well. In keras.js, convolution is implemented with the oft-used im2col transformation to turn it into
55.
Keras.js – Run trained Keras models in your browser(github.com)
192 points by transcranial 10y ago|25 comments
56.
by transcranial 10y ago
If anybody likes transparent backgrounds, I wrote up a quick guide on how to enable it for Atom: https://github.com/transcranial/atom-transparency
57.
by transcranial 10y ago
And https://arxiv.org/abs/1603.07285 for the corresponding paper. Really clear and easy-to-understand explanation of some of the math.
58.
by transcranial 10y ago
Interesting that the base data consists of URLs. I guess it makes sense given copyright issues. Anybody know what the ballpark expected half-life of such URLs?
59.
by transcranial 10y ago
Agreed! In many ways it is massively underrated precisely because it isn't backed by a Google or Facebook. Definitely glad that people are seeing its awesomeness based on merits. I love React and its whole ecosystem, but Vue definitely
60.
Microsoft creates 5000-person AI division(news.microsoft.com)
71 points by transcranial 10y ago|15 comments

More ›