Hacker News Reader

projectorlochsaon Deep Learning’s Impact on Image Processing, Mathematics, and Humanity

Deep learning based style transfer that got viral a year or two ago was worse than this 2008 result, especially worse for that hone application to portraits.

The hype was nowhere near just because there's no deepness.

I admit that https://arxiv.org/abs/1705.01088 this beats everything and is extremely powerful and simple but the hype for deep things seems a bit too strong.

projectorlochsaon Deep Learning’s Impact on Image Processing, Mathematics, and Humanity

Author mentions style transfer as a product of NN research.

Yet this paper [1] from 2014 works extremely well, no neural networks. Here's one from 2008 [2] that worked extremely well too, better than most early NN techniques.

Although, deep learning is far from the intuitive approach that existed 10 years ago. It's clearer now how to reason about models, layers, activation functions etc. As much as there's no mathematical foundation, I do not believe it helps that much in the case of SVMs or whatever else. There still has to be experimentation, approximation and proper testing.

[1]: https://dspace.mit.edu/handle/1721.1/100018 [2]: http://link.springer.com/chapter/10.1007%2F978-3-540-88690-7...

projectorlochsaon Juice may taste good, but unlike the fruit, it’s not healthy

Concept of an ingredient being healthy does not exist in science. Is diet filled with fruits and vegetables healthy? Of course. Starchy foods with high nutrients per calories ratio have super high unattainable overdose amount. That's why they are safe.

Eating 3000 kcal of bananas metabolises quite differently than 3000 kcal of glucose sugar.

projectorlochsaon Chinese Offer to Eat Denmark’s Oyster Problem

Consuming oysters in a larger than usual amounts would result in high amounts of unwanted metals from the polluted sea (mercury etc.) in the body. Oysters are wonderful sea cleaners.

It is better for a larger population to consume them all, instead of a smaller one consuming more per individual.

projectorlochsaon Limits of computation

I remember watching some documentaries and reading some articles, and remember Hugo de Garis talking about femtotech. I thought he was a pioneer in the technology, seems he's just a bystander.

He and Ben Goertzel seem to look as pioneers but never produce something tangible.

projectorlochsaon You Can’t Trust What You Read About Nutrition

There aren't million other factors. Saying that cholesterol is affected only by saturated fat is not true. just like saying it doesn't affect it at all.

there is no coherent nutrition advice available at the moment, other than observational evidence of various diseases forming when a population is fed an improper diet.

for example, if starchivores like apes are given huge amounts of sugary foods, they will develop atherosclerosis, the same happens if they're given huge amounts of fatty foods. if carnivores were fed huge amounts of fatty foods, they don't get atherosclerosis, only if you do not remove their thyroid gland, if they on the other hand are fed sugary foods, they develop atherosclerosis. so the mechanism involved in absorbing fat and using it (through various types of cholesterol - lipoproteins), seems to be different between the herbivores, omnivores, carnivores. interesting thing about this is that the discrete classification is faulty and that different hormonal balance can give rise to no or different disease.

there's plenty of examples like this one. but it has nothing to do with the individual as population statistic is quite different from needs of an individual.

projectorlochsaon Alphabet's Self-Driving Cars to Get Their First Real Riders

The main approach of how AlphaGo went around the issue of never seeing the positions and playing good moves is search (monte carlo tree search).

In the same way, if the world is modelled accurately and succintly, can a car predict the various scenarios and adapt, even if the current sitation was not seen before (extremely high speeds or unusual number of cars, etc.)

projectorlochsaon You Can’t Trust What You Read About Nutrition

Not really "incorrect". Dietary cholesterol does effect cholesterol amounts in the body, in addition to other food intake, hormones etc.

projectorlochsaon Remembering the World’s Oldest Person, in the Objects She Left Behind

Surprisingly, many of them are regular smokers. If your genes treat you well it seems diet is negligible.

For example, there are genetic abnormalities resulting in extremely low cholesterol, no matter how much calories they eat, atherosclerosis will never appear.

Living long is obviously a sign of serious immunity towards all kinds of poisons.

projectorlochsaon Atmospheric CO2 levels accelerate upwards

Deforestation of Argentinian and Brazilian agriculture tops that any day.

Phytoplankton loss does that too.

projectorlochsaon What is this colored fiber in my chicken?

In absolute numbers there are more slaves today than in any moment in history of humanity.

Speciesist and anthropocentric mentality that is reinforced by world's most popular ideologies - shallow postmodernism, christianity and other religious ideologies - is what keeps this stuff alive.

It is a shame war is necessary to force people to abandon congitive dissonance and start making a better place of this world. At least we in first world countries have all the time to make the change and be the role model, but we seem to be doing business-as-usual, at least the majority of us.

projectorlochsaon Paradoxes of probability and other statistical strangeness

I think it is impossible James-Stein estimator of the whole sample can outperform all of the three MLE for each particular. Not that I would know of a way to generate a particular from a joint estimator.

projectorlochsaon Paradoxes of probability and other statistical strangeness

Well this does make a little bit of sense if after estimating things get worse for a particular case.

If MSEs were 5, 5, 100, and after they are 10, 10, 80, the total got better but the prediction for the other two got worse.

projectorlochsaon Milk that lasts for months

More than 85% of soy grown is for non-human animal consumption. Animal agriculture in Brazil is the leading cause of Amazon deforestation, likewise, animal agriculture in Argentina.

I'm not even sure if you've read the source you've given as it states the same.

Not to mention that most soy humans consume is non-GMO (which excludes practically everything from Argentina and Brazil), and that soy is grown in China, and for example in suprisingly large quantities in Serbia. So it's not even necessary to cut rainforests.

projectorlochsaon Milk that lasts for months

Clinical studies show no effects of soy protein or isoflavones on reproductive hormones in men: results of a meta-analysis.

http://www.fertstert.org/article/S0015-0282(09)00966-2/fullt...

Good thing we've got science to reinforce our claims.

Not to mention the fact that you're drinking milk from a lactating female filled with animal estrogen yet you worry about plant estrogen.

projectorlochsaon Milk that lasts for months

http://www.perfectdayfoods.com/

There's also milk made with casein protein and all the goodies from original cow milk, but without the cows.

Also has a fairly long shelf life.

projectorlochsaon Milk that lasts for months

Another milk that lasts months is soy milk. Nutritionally superior and cruelty free!

projectorlochsaon The Case Against Eating Fish

What argument, who's arguing? I'm saying that making decisions for your own particular situation based on what would happen to population statistics is fairly tricky.

Especially because of the epidemiological nature of these studies, lack of interventional studies for this particular subject and a bunch of other completely logical reasons.

projectorlochsaon The Case Against Eating Fish

The problem is that one individual is not an average population. Average population in USA is incapable of eating right.

There's a bunch of stuff that other than fish that would have the same if not better effect and it has no side-effects.

projectorlochsaon LaTeX Tooling Guide

You haven't really explained why it doesn't.

Choosing a language allows you to automatically translate a bunch of strings used by various packages.

For examples Chapter, Section, TOC and everything that is autogenerated will be translated to a language. It has nothing to do with utf8.

You can remove babel and Chinese, German, Swedish can work in the same file regardless.

\usepackage{noto}

What's hard about that?

projectorlochsaon LaTeX Tooling Guide

\usepackage[utf8]{inputenc}

\usepackage[LANG_OF_CHOICE]{babel}

solves that problem.

there's plenty of font packages available for pdflatex.

I use newtxmath and newtxtext and libertine math and text all the time.

projectorlochsaon Building Safe AI: A Tutorial on Homomorphically Encrypted Deep Learning

Deep learning automatic driving systems that want to work have to have world models.

It's Automatic Driving 101.

These models don't have to be as explicit as formulas but can be approximations of reality through beam search (having multiple steering hypotheses at once and then picking the most likely one etc.), model ensembles, some bayesian state exploration or anything that isn't random search.

projectorlochsaon Startup Serves Up Chicken Produced From Cells in Lab

Same thing a person grown on vegetables and fruits might say. Thousands of choices, all different in textures and tastes and can be combined into extremely delicious meals.

The quality of life is subjective and taste can easily be acquired in a way that one can continue feeling satisfied with the foods they eat.

What you might need is a tastebud reset, which can easily be accomplished and after that you'll find anything extremely delicious.

You mentioned that you've just tried an ingredient and didn't like it which isn't even comparable to your dietary regime which was filled with meat during an extremely long period of time. To acquire taste takes effort or some out of the ordinary measures (a week or two of one ingredient diet to reset the taste buds, or that alchohol trick repetatively etc.).

projectorlochsaon Startup Serves Up Chicken Produced From Cells in Lab

Taste is acquired, it is not in your genes. Of course you will find foods and tastes that were present through your whole life better than new textures and tastes. It's normal.

Eventually one either gets used and enjoys the new foods or makes tasty meals in combination.

For me, getting drunk and then eating the disgusting thing in the morning made me love it. It's undescribable how delicious things are when one is dehydrated and when the disgusting taste of beer was the only thing in the mouth the night before.

projectorlochsaon What fossil fuels are doing to our oceans

Data centers are about 1-2% of emissions.

Animal agriculture is comparable to all transit, while heating/cooling/electricity is heavily at the top.

If I were profiling I'd optimize the highest bottleneck first.

projectorlochsaon What fossil fuels are doing to our oceans

Although, male calves are killed in milk production. Dairy cows cannot produce thousands of liters of milk without living a short life. Chickens that produce eggs were bred into monsters that can't live long.

I do not think any hate is rational and believe that any kind of vegan propaganda is beneficial to the cause.

Fossil fuels are nothing compared to nitrite waste.

projectorlochsaon Prophet: forecasting at scale

Probabilistic graphical models.

That's the foundation. The way you set up your model is by nodes and edges that specify the flow of influence (directed or undirected). Then it seems that there are general methods for inference and learning on any kind of graph one might pose.

For simple graphs (and simple is something one might want when modelling) the methods should be fairly effective.

Unfortunately, the biggest book on the subject that I know (Koller & Friedman) isn't accessible. Koller's course is also not that accessible.

projectorlochsaon People who eat 10 portions of fruit and vegetables a day

What does acid forming mean?

projectorlochsaon Susan Fowler fears she's now the target of a smear campaign

N that the GP is talking about is number of vehicles and number of requests.

Yes, service can obviously work with subpar algorithms but to really succeed at pricing it as cheaply as possible it requires practically an ability to successfully predict the whole day and then optimizing on the NP-hard problem of that whole day. Maybe sampling a million day variants while routing to make a single decision (which driver should pickup the next request).

Of course brute force greedy algorithms work but they can be, on a hundred vehicle scale, 30% away from the optimum cost.

I've been downvoted to oblivion so I cannot longer keep participating in this discussion (HN will shadowban me).

projectorlochsaon Susan Fowler fears she's now the target of a smear campaign

Yeah, problem is so simple that they have job listings for AI research engineers that should have a little bit of experience in TensorFlow, Theano, Caffe or Torch (obviously some lovely deep learning on who knows what). About a year ago they were searching for PhDs and asked for experience with combinatorial algorithms, especially travelling salesman and vehicle routing problems.