lars

7 ms·

31.
by lars 7y ago
Ok. I can give you a 100 million floats, and tell you that if you multiply, add and threshold them in a certain order against a set of pixels, that tells you if the picture is of a cat or a dog. That is an exact reproducable procedure that
32.
by lars 7y ago
If I show you a picture and ask you if its a dog or a cat, you can answer the question, but you can't tell me why. You can make up a story to justify why it's a dog or a cat, but if someone tries to implement the procedure you giv
33.
by lars 7y ago
He's not wrong though. We know precisely how any given neural network works, it's just not understandable to us. The point is though, the functions we're trying to learn, like image classifiers, have no responsibility to us t
34.
by lars 7y ago
While the sources you mention are all good, I don't think any of them actually have much to say about what the link is about, which is bayesian deep learning.
35.
by lars 7y ago
Friendly reminder that churn rate isn't a meaningful metric, because churn probability is not constant throughout a customer lifetime: https://medium.com/swlh/youre-all-calculating-churn-rates-wr... The churn rate
36.
by lars 7y ago
The react license had "we won't sue you for violating our patents, except if you sue us first.", which entails "if you use this, you can't sue us for violating your patents, because by using this you guarantee that
37.
by lars 7y ago
These ML patents are damaging the progress of AI research, and also the commercial use of this intellectual progress. One way to attack this would be if researchers outside of Google and Facebook started patenting their discoveries, but the
38.
by lars 7y ago
A hash table can also approximate any function, so the "universal function approximation" thing is a bit oversold. It isn't really what matters. What matters is how well methods generalize beyond the training data, and how mu
39.
You’re all calculating churn rates wrong(medium.com)
3 points by lars 8y ago|0 comments
40.
We ran a recommender algorithm over US Congress voting data(metricspace.ai)
4 points by lars 8y ago|0 comments
41.
by lars 8y ago
Amount of peace and wealth are clearly not fully predictive of good mental health. Public health officials in England expected a civilian mental health crisis during WW2. This didn't happen, the incidence of "neurosis" remain
42.
by lars 8y ago
Maximizing total synapse strength would as useful as maximizing the number of 1-bits on your hard drive. It's not what you want :)
43.
by lars 8y ago
This was Adrian Thompson using genetic algorithms to design FPGAs: https://www.damninteresting.com/on-the-origin-of-circuits/
44.
by lars 9y ago
Only 39% of the submissions to the latest ICLR (machine learning conference) were from .edu emails [1]. There's a lot of research going on outside of academia, and there's also a large portion of academia that doesn't use the
45.
by lars 9y ago
I don't want to promote anyone in particular, but there are a few vendors here, for example: https://www.ovh.ie/dedicated_servers/gpu/ https://www.exoscale.ch/gpu/ http://www.
46.
by lars 9y ago
Oh, there are smaller vendors renting out Pascal architecture GPUs. Yet Amazon and Azure are stuck at prev-gen (M60), while Google is all the way back at prev-prev-gen (K80).
47.
by lars 9y ago
They say "next-generation", but these are M60 GPUs, which are very much "previous-generation". Current generation would be P100 GPUs. I am in the market for a cloud GPU offering, and I have to say the big cloud providers
48.
StackGAN: Photo-Realistic Image Synthesis with Generative Adversarial Networks(arxiv.org)
2 points by lars 10y ago|0 comments
49.
IBM: Mind reading is less than five years away. For real. (2011)(cnet.com)
2 points by lars 10y ago|1 comments
50.
by lars 10y ago
I've been going to US tech conferences yearly in the last few years. As of 2016 this has stopped, since the fact that I was snowboarding in Iran in 2011 now means I can't get a visa through ESTA. I now have to take a day off work,
51.
Recurrent Highway Networks(arxiv.org)
1 points by lars 10y ago|0 comments
52.
by lars 10y ago
> One main thing it lacks is imagination. [...] For example, if I ask you to imagine a Guitar playing Dolphin, you could imagine it and even recognize it from a cartoon even though you have never seen it in your life before. Not so for D
53.
by lars 10y ago
I agree that scale is an important factor in deep learning's success, but that Google experiment ended up being a good example of how not to do it. They used 16000 CPU cores to get that cat detector. A short while later, a group at Bai
54.
by lars 10y ago
Go to "speech, inking & typing" settings. The dialog literally says they will collect your typing history. The traffic is encrypted, and no one has been able to MITM it yet, as far as I know. But Microsoft has confirmed in the
55.
by lars 10y ago
More than that, what it installs fits any classical description of malware. Windows 10 comes with an OS wide key logger, that logs what you write in any application, and sends it to Microsoft. (This usually evokes disbelief. If you think I&
56.
by lars 10y ago
Wow, I'm glad this was posted. I think the way we use technology today will be looked back on the same way we look back at naive cigarette smoking in the 1950s. Modern app design isn't about creating things that are good for the u
57.
by lars 10y ago
https://larseidnes.com/2015/10/13/auto-generating-clickbait-... http://clickotron.com
58.
by lars 11y ago
This seems a little myopic. What distinguishes basic income from the welfare systems implemented in, say, northern Europe? Seems like there are many countries that have had basic income for decades. Couldn't you look to them and have m
59.
Painting Video with Neural Networks(larseidnes.com)
2 points by lars 11y ago|0 comments
60.
by lars 11y ago
Spoilers: Women working on pedal driven sewing machines were complaining of back pains. An alert medical professional also discovered that the women were getting off by the leg movement that drove the machine. Clearly, it was this immoral b

More ›