pps43

47 ms·

361.
by pps43 10y ago
I'm not sure about human players, but it might be possible to build an adversarial neural network that beats AlphaGo (and only AlphaGo).
362.
by pps43 10y ago
Accuracy from one quadrant and four quadrants would be the same, though, and full circle is easier to code (no extra conditions). To get better accuracy you would need to cut out some parts that are clearly inside or outside the circle, and
363.
by pps43 10y ago
You can try the same with quasi-random rather than pseudo-random numbers (Sobol, Halton, or even just additive). Should probably get better accuracy with the same number of iterations.
364.
by pps43 10y ago
The section on the chess game is not particularly persuasive. According to Stockfish, black already has winning position. So Frank (playing white) can no longer make mistakes. All moves at his disposal lead to the same outcome. "Queen
365.
by pps43 10y ago
> if I use "aaaaaaaaa" as a password on a website I know full well what I'm doing. There's no substitute for length. "dog" is a bad password. "doooooooooooooooog" is much better. The entropy is sti
366.
by pps43 10y ago
> I seem to remember a github repo that was posted to HN a couple of years back that did exactly that. They showed the same data set and presented 3 different ways with 3 entirely different conclusions. Probably this one: http:/&#x
367.
by pps43 10y ago
Key word here is "potentially". There's not much oil production in Black Sea. There are much better oil fields in Caspian sea nearby (which the linked article even mentions), let alone Siberia.
368.
by pps43 10y ago
> $25 an hour This is counterproductive. I can (and often do) spend an hour to answer a question on StackOverflow, for free. Offer me $25/hour for answering StackOverflow questions and I'll decline - it's well below my hou
369.
by pps43 10y ago
There are no strategic resources to speak of in Crimea. Black Sea fleet has no strategic importance for Russia because it is locked by Turkey, still a NATO country. All strategic ports are in the North and on the Pacific. It was more like &
370.
by pps43 10y ago
Not really. In the first two years (from May 1957 to May 1959), R-7 had 56% failure rate (23 launches, 13 failures). It is now one of the most reliable rockets.
371.
by pps43 10y ago
> invading North Korea in 2003 would have been a far better choice This sounds just like the rationale for the Vietnam war or the recent annexation of Crimea.
372.
by pps43 10y ago
Yes. The credit bureau just shows the balance. Card issuer knows whether you are a transactor (pay the entire balance each month) or revolver (carry over some of the balance and therefore pay interest on it). This is a strong risk predictor
373.
by pps43 10y ago
If you're not doing anything wrong you have nothing to hide?
374.
by pps43 10y ago
Not really. With over-fitting, performance goes down, not just plateaues.
375.
by pps43 10y ago
I suspect OTA update was applied simultaneously rather than staggered over the course of the year, so probably not. You can control for seasonality using some other data as a proxy, but the more adjustments you make, the more potential ther
376.
by pps43 10y ago
Confounding factors. For example, crashes can be more likely after sunset or during inclement weather so seasonality can play a role. Or maybe the drivers who had Tesla longer got used to it.
377.
by pps43 10y ago
The question was not whether I'm a native English speaker (obviously not), but whether I was educated in the US. I believe the latter is a much harder question.
378.
by pps43 10y ago
Correct on both counts, although it's not quite clear how much of the difference is from place of education rather than native language. Math is the same no matter where you study it, at least in theory. Probably need a bigger sample.
379.
by pps43 10y ago
Interesting. Any telltale signs? I'm assuming it's not the difference between native English speaker and ESL, but specific words or grammar patterns? Can you tell from my comment history whether I was educated in the US?
380.
by pps43 10y ago
Someone on H1B usually is also waiting for employment-based green card or promise to file for one. Going to another employer means that this multi-year process needs to be restarted.
381.
by pps43 10y ago
How long will it take to "train somebody" if you need to translate technical documentation from English to Japanese? Some of foreign countries have free education. Why not hire their best graduates? They don't have to pay bac
382.
by pps43 10y ago
The key point is self-learning, or ability of AI to build AI that's better, if only a little. This is different from, say, AlphaGo playing against itself to train its neural network - we want AI 1.0 to write AI 2.0, not just tweak so
383.
by pps43 10y ago
> I want to build a simple circuit with a few LEDs and understand completely how it works and why it was designed that way. Once you get bored by "battery plus LED" circuits, I recommend Nibbler 4 bit CPU [1] and transistor clo
384.
by pps43 10y ago
> your insurance rates would go down if all your health info were made public and the insurance provider was allowed to factor that in, right? It's already happening, the only difference being that it's the people rather than i
385.
by pps43 10y ago
> anyone who does it deserves a swift kick to the nuts. Let's say you want to buy a house and apply for mortgage. The bank can either look at your credit history to determine your probability of default, or not look at your credit h
386.
by pps43 10y ago
> No clue how they got all that shit. Credit bureaus and companies like Axciom or DataTree collect, keep and sell tons of data. At the very least you should request your credit report (it's free if you don't want FICO score) an
387.
by pps43 10y ago
What if you do the same calculation for Indian Point? I suspect the cost of relocating everyone in New York City and boroughs would be pretty substantial.
388.
by pps43 10y ago
Processing power is not a problem these days. Maybe if you feed all channels at ultrasound frequency to ADC and then try to do all processing in software. But you don't have to, it's oftentimes easier to have hardware (FPGA) pseud
389.
by pps43 10y ago
The reasons ultrasound machines are expensive are not technical, it's mostly low volume and high overhead to comply with all rules and regulations. You can make a useful ultrasound machine with just one transducer (e.g., to measure blo
390.
by pps43 10y ago
API is Encrypt(data,key) and Decrypt(data,key). There's no IV. It's just a block cipher. Once you start adding things like CTR mode, MAC, etc. the API is no longer as simple.

More ›