yonkshi

5 ms·

1.
by yonkshi 7y ago
Unfortunately any AI systems will encounter the verification dilemma, the more powerful an AI system becomes, the less verifiable it would be.
2.
by yonkshi 7y ago
I agree. I think a big part of this problem is that smaller companies usually cannot afford AI research. I would even go as far as to say there are more AI companies than capable AI researchers, and this causes a large number of faux-AI com
3.
by yonkshi 7y ago
AGI is a gradient, not an arbitrary threshold. We are not capable of recreating human level intelligence yet, but our modern algorithms had become magnitudes better at generalization and sample efficiency. And this trend is not showing any
4.
by yonkshi 7y ago
My guess is to prevent ad-fraud by website owners. It's a lot harder to detect fraudulent clicks/impressions if all data are routed through website. The cost of potentially blocked by ad blockers is finite (A percentage of total r
5.
by yonkshi 7y ago
Not only dupe, Forbes also has the most click baity title amongst them all. Forbes is like buzzfeed now
6.
by yonkshi 7y ago
It appears they are using an algorithmic process to perform the heartbeat trading, and they are patentable[0] as long as they are not illegal. Tax loopholes are not illegal but very unethical, I think the policy makers share equal if not mo
7.
by yonkshi 7y ago
This is such a cool concept, I really wish these guys would succeed. Though last time I heard they gave up developing their own rocket (as a second stage to space). Anyone who's in the loop, does Stratolaunch currently have any potenti
8.
by yonkshi 7y ago
Elaborate? We are discussing event probability, not logical inference, so how can there be fallacy in the first place.
9.
by yonkshi 7y ago
Occam's razor applies here. Your theory is that she had 1. intended to assassinate Assange, 2. attempted to cover it as a joke and then 3. further covering it up by implying she doesn't remember this statement. Let me offer you an
10.
by yonkshi 7y ago
My friend Gibbs invented this really efficient way to learn.
11.
by yonkshi 7y ago
More specifically Markov chain with monte carlo method (MCMC)
12.
by yonkshi 7y ago
Sure theoretical advances may be evolutionary, but they resulted in exponential reduction of parameter space and sample complexity. These advances outpace Moore's law by a large margin. Hardware served as a catalyst, but it was not a n
13.
by yonkshi 8y ago
Less than four years ago our state-of-the-art RL system (DQN) could only beat some Atari games. Now we can almost beat best human player in SC. That to me is very impressive.
14.
by yonkshi 8y ago
Currently visiting China for Chinese New Year, and can't help but to notice the anti-US anti-Apple sentiment that arose recently. Although anecdotal, I think the nationalistic sentiment has a lot to do with the dip in sales. To put in
15.
by yonkshi 8y ago
..could you elaborate? Are you saying ML is dangerous because it’s based on mathematics, and math is too clean to handle real world data?
16.
by yonkshi 8y ago
Sure. I don't have any holistic survey to prove my point, but an example of recent progress in terms of sample efficiency is this paper[0]. Derivatives of this paper have been used to solve Sudoku[1], Starcraft II[2] and more [3]. This
17.
by yonkshi 8y ago
The sample efficiency of ML systems are increasing rapidly in the theoretical realm, though the available data are growing magnitudes faster than the progress in ML sample efficiency. The article has a valid point: It's far cheaper to
18.
by yonkshi 8y ago
Are you suggesting distributed NN training through user's browsers? This is a very sketchy area along the lines of browser based bitcoin mining. You don't want to run any computation heavy code on client side without user's k
19.
by yonkshi 8y ago
There is currently no NN learning algorithm that can handle massively paralleled training. We can use some simple fixes such as mean gradient but they have severe limitations, and their limitation grows as you scale up. Currently even with
20.
by yonkshi 8y ago
JS/V8 may beat Python in general operations but definitely not in math specialized libraries. A lot of python packages support OpenBLAS and AVX/FMA by default which gives huge boosts to math/tensor operations. Correct me if I
21.
by yonkshi 8y ago
I think we will also see proportionally more and more chips coming out of US companies as well. When Google started their own chips (TPUs), they just kickstarted a vertical integration race amongst the tech giants. Apple, Intel, Nvidia were
22.
by yonkshi 8y ago
I'd be interested as well, I've written a few private scrapers for *A/OW award search, currently have small scale MS going on, would love to connect with more fellow MS-ers. my email is nopctoday at gmail.
23.
by yonkshi 8y ago
Correct me if I'm wrong, but I think the significance of this finding is not so much about temperature, rather it's a big piece of the puzzle in understanding superconductivity. This is the first time that we are able to achieve s
24.
by yonkshi 8y ago
And Apple's own statement said that the journalist is probably confusion a separate, software based, incident with this incident. https://www.apple.com/newsroom/2018/10/what-businessweek-got...
25.
by yonkshi 8y ago
I don't think most ML experts would agree with that, a big reason DL became popular are the huge improvements they brought to CV and NLP fields. In many ways, traditional approaches were harder because you need huge amount of domain ex
26.
by yonkshi 8y ago
I spent 10 minutes on your site, still have no idea what it's about. If you say it's the "decentralized facebook", I don't see anything about a feed/updates, no way to post/broadcast anything. The only thi
27.
by yonkshi 8y ago
Not necessarily. E.g. making Ramen noodles, much cheaper to go to a ramen shop than to attempt to make it yourself. Economy of scale.
28.
by yonkshi 8y ago
AI can also be used to identify these fake videos, and they are probably better than human identification. There will be a rise of AI forensics. The type of neural network they use (GAN) works by having two networks battling each other, one
29.
by yonkshi 8y ago
I had a discussion with a buddy about the California high speed rail and similar land prices issue. Surprisingly, eminent domain allows the gov to acquire land fairly cheaply. https://en.wikipedia.org/wiki/Eminent_domai
30.
by yonkshi 8y ago
I advise you to reflect on the bias of your own view points. China has been pulling the same bullying tactics towards EU and S.Korea/Japan as well. IP transfer, forced partnership, etc. Notice how EU, South Korea and Japan haven't

More ›