argonaut

29 ms·

271.
by argonaut 9y ago
This is essentially an argument against having a minimum wage in general (and more broadly, certain labor regulations). It's a complex issue that we could argue about for pages and pages of comments, but suffice to say the vast majorit
272.
by argonaut 9y ago
Yes, the network has picked out some image features. But those image features are being trained towards outputing what are random predictions, so the features are essentially noise.
273.
by argonaut 9y ago
For large dimensions, a normal linear search.
274.
by argonaut 9y ago
I you would downsize the frames first and work on multiple frames in parallel (e.g. however many nets I can fit in VRAM). I find it hard to believe that it wouldn't work after these things.
275.
by argonaut 9y ago
Self-driving is not going to be solved by throwing money at it. If it was, Google would have "solved" it already and Apple wouldn't have abandoned their effort. The obvious shortage is engineering talent. This is more debatab
276.
by argonaut 9y ago
That's news to me! I'm pretty darn sure the Titan X has no problems with inference on the amounts of data that a single user would want inference for.
277.
by argonaut 9y ago
It's not progress if the market leader, Google, already knew (or at least decided) that approach was wrong.
278.
by argonaut 9y ago
Yes, it's all bogus marketing, but people mocking Juicero should be reminded that bottled water and luxury fashion goods of any kind are all multi-hundred billion dollar markets, not to mention any of the trillion (?) dollar commodity
279.
by argonaut 10y ago
Why is having a huge cluster of GPUs/CPUs so important for AlphaGo? We humans only using about 100 Watts of power. We don't need megawatts of datacenter power. Visual SLAM is an open research problem, and lidar lets companies side
280.
by argonaut 10y ago
No, generally models are higher dimensional than their inputs (AKA deep learning models generally have more parameters than the number of features).
281.
by argonaut 10y ago
This is speculative, but it seems like the privacy aspect is oversold as it may be possible to reverse engineer the input data from the model updates. The point is that the model updates themselves are specific to each user.
282.
by argonaut 10y ago
My ML background is primarily undergraduate work, and even then it was extremely common.
283.
by argonaut 10y ago
Saying you are 95% confident AGI will come between 50-100 years (a "huge" confidence interval), is still far too confident.
284.
by argonaut 10y ago
You won't find a list of college with legacy admissions because almost all private colleges do it.
285.
by argonaut 10y ago
You admit that you don't know what the actual probabilities are. Or you admit you have an extremely low confidence in your estimate of the probabilities.
286.
by argonaut 10y ago
Only the first reason is correct (NNs are good at data with dimensional relationships). The second reason is pretty bogus (text/images more graspable). It's valid if you're talking about mass media / popular press. But f
287.
by argonaut 10y ago
The Deep PCA paper shows no such thing. They've cherry-picked bad CNN baselines. The red flag for me is that the "state of the art" methods they compare to achieve > 1% test error on vanilla MNIST, which is totally wrong s
288.
by argonaut 10y ago
Not sure where you're getting this from, but CNNs are definitely still state of the art on MNIST. Papers often cite outdated CNN numbers; in fact the best published CNN accuracy numbers are probably lower than they could be - vanilla M
289.
by argonaut 10y ago
The fact that this is Tweetdeck makes me believe this isn't a core initiative by Twitter's leadership, but probably a side project initiated by some individual contributors at Twitter. So this shouldn't be taken to be some bi
290.
by argonaut 10y ago
"It's not really a neural network" is an understatement. Perceptrons are about as related to neural nets as linear regression is (note that you can "train" linear regression with stochastic gradient descent).
291.
by argonaut 10y ago
Your history is a little mixed up. LeNet-5 had 7 layers, for example (late 1990s). Regardless, neural nets fell out of favor because they were seen as overcomplicated and even though they achieved competitive accuracy (yes, even when they w
292.
by argonaut 10y ago
Rosenblatt's perceptron has little to do with neural nets. Geoffrey Hinton regrets coining the name "multi-layer perceptron" precisely because they're really unrelated.
293.
by argonaut 10y ago
Yes. I'm not sure what point you're trying to make about tautology. Adversarial ML examples are clearly errors if they were part of the test set, and part of "correctness" is reducing the error rate (correct model, corre
294.
by argonaut 10y ago
I think it's safe to say it's way too early to call it a "hit." Remember that Google Glasses were also a "hit" among bloggers and techies and other influential people. Of course in hindsight we know this was th
295.
by argonaut 10y ago
Opsec safety is not the biggest problem with autonomous driving. It is a secondary problem at best and one that can be addressed (though certainly there is room for improvement) using normal security techniques. Correctness is the biggest
296.
by argonaut 10y ago
Still doesn't make sense. You don't crochet by finding a weave that's compatible with your personality or whatever. You crochet by following correct technique and stitching.
297.
by argonaut 10y ago
Autonomous driving is a software engineering/research problem, not a hardware bottleneck. You couldn't train a good autonomous driving agent even if you used a massive GPU cluster. I'm not saying better hardware doesn't
298.
by argonaut 10y ago
Mark Zuckerberg is a particularly bad example, since he was the son of two white-collar professionals, went to an elite private school (Phillips Exeter) for two years, and then went to Harvard. He was probably eligible for little to no fina
299.
by argonaut 10y ago
How much do you make? You're probably in the top 10% of household income (>= 140k).
300.
AI’s PR Problem(technologyreview.com)
3 points by argonaut 10y ago|0 comments

More ›