Jump to content

Wikipedia:Reference desk/Archives/Science/2017 May 31

From Wikipedia, the free encyclopedia
Science desk
< May 30 << Apr | May | Jun >> June 1 >
Welcome to the Wikipedia Science Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.


May 31

[edit]

Bbits per second

[edit]

Can you calculate the rate of bits per second at which a human learn? — Preceding unsigned comment added by 31.177.96.46 (talk) 15:44, 31 May 2017 (UTC)[reply]

@31.177.96.46: No, because this is not a well-defined quantity (yet). For example, how many bits of information is learning an Axel jump and at what point in time does it become fully "learned"? We can't answer the former and the latter does not admit a natural definition.--Jasper Deng (talk) 16:01, 31 May 2017 (UTC)[reply]
Since even world-class figure skaters fall on jumps, the question of when they "fully" learn it might be answered "never". ←Baseball Bugs What's up, Doc? carrots20:47, 31 May 2017 (UTC)[reply]
And yet you can safely assume that most senior-level skaters have "learned" it in the everyday sense of "learned", even if they cannot do it 100%, because they know all the different components of the move and (usually) can demonstrate it to others. Hence the point remains that there is no natural definition of "learned" suitable for a calculation like the OP asks for. Also, to be a bit clearer, my statement about "learned" is intended specifically for each bit of information. Even if we could model a learning process as a stream of bits entering the mind, at what point does each bit become learned? "Learned" does not obey the law of the excluded middle.--Jasper Deng (talk) 21:21, 31 May 2017 (UTC)[reply]
Figure skating is fun, but this discussion is distracting from the pedagogical purpose of bringing it up.
Doing back-to-back jumps and having to skate several minutes are handicaps. A 10 second routine with 1 jump would be easier and even female skaters seem to do double jumps like they're a piece of cake. Maybe some human has in fact become practically immune to falling from a single Axel (assuming fair conditions. Not drunk, tired etc, no wind, earthquakes but no mulligans). No one will ever do a centuple Axel without cheating. Sagittarian Milky Way (talk) 21:41, 31 May 2017 (UTC)[reply]
@Sagittarian Milky Way: Next time, make sure you don't overrun my comment! Also, not all jumps are strictly done back-to-back. When it is, we call it a combination.--Jasper Deng (talk) 23:32, 31 May 2017 (UTC)[reply]
Human senses and the human brain, or Neurons in general are far beyond binary math. If you would try to brake this down into the binary system you could count all the nerve cells and multiply the number with 256 kbps, which is known as near real "resolution" in mp3-audio. On one hand the result would probably be much to high but on the other your Question is alike much to wrong. --Kharon (talk) 20:10, 31 May 2017 (UTC)[reply]
You can, however set an upper limit by calculating the rate at which the human retina transmits data to the brain (around 10 million bits per second, which is close to an Ethernet connection) and repeating for hearing, touch, etc.[1][2] You can't learn any faster than that, but you can learn a lot slower. Judging by some of the people I come in contact with, learning can sometimes be as slow as several bits per year, or can even go negative. --Guy Macon (talk) 20:25, 31 May 2017 (UTC)[reply]
  • This article deserves to be shared every time this question comes up. The OP should read it daily until it becomes ingrained. --Jayron32 23:35, 31 May 2017 (UTC)[reply]
    • The bit is a unit of information, independent of the machine acting on it. So the OP's question is quite valid to ask even without the proposition that we're computers. Ultimately, the state of our mind at a given moment can be given as a large but finite amount of values of state variables, so information theory still applies.--Jasper Deng (talk) 23:38, 31 May 2017 (UTC)[reply]
    • Let's not forget that the article linked by Jayron is just an opinion about how things work or could work. And it's a minority view in cognitive science.--Hofhof (talk) 12:35, 1 June 2017 (UTC)[reply]
    • This article provides no evidence for the very big claims it makes. It claims that humans are not born with and never develop representations, rules, models, algorithms or memories which is a ridiculous claim alone. He cites not basis for the claim that there is no symbolic representation of the world in the brain and fails to offer a clear alternative. 208.90.213.186 (talk) 20:14, 1 June 2017 (UTC)[reply]
      • It also commits a crucial error, namely saying that a byte is 64 bits. The article on the byte does not suggest that 64 bits was ever considered a "byte" and in common computer science parlance, a byte is always 8 bits.--Jasper Deng (talk) 04:50, 2 June 2017 (UTC)[reply]

Thomas Landauer, in this paper, came up with an estimate that the total amount of information stored by a human brain was on the order of one gigabyte, and the learning rate for some tasks he measured was about 1 bit per second. I don't know how well this paper has stood the test of time, but a look at the 213 papers that cite it might be informative. Mike Christie (talk - contribs - library) 09:09, 1 June 2017 (UTC)[reply]

  • This paper suffers from the crucial flaw I mentioned above: the definition of "learned" is not specified and this analysis cannot be applied to learned activities such as figure skating.--Jasper Deng (talk) 04:50, 2 June 2017 (UTC)[reply]
The world record for memorising random binary digits is 2385 digits memorised in half an hour. While agreeing that the exact definition of "learning" is undefined in this whole discussion, this suggests that a rate of 1 bit per second is in the right ball park. Gandalf61 (talk) 11:45, 2 June 2017 (UTC)[reply]
But if a random word from the thousand most common is shown for 1 second and this is done twice without intermission it would take 20 bits to store that data but most people could remember that. Even if English could be compressed to about 1 bit per letter you could make the person also remember the color of the words and color them one of 10 colors and if he can remember that it'd be about a thousand bits in 2 seconds. And Kim Peek memorized the contents of at least 12,000 books and had barely a billion waking seconds to do it in so that seems more like 10 bits per second. Sagittarian Milky Way (talk) 22:39, 2 June 2017 (UTC)[reply]
It might be true that humans have that "learning" rate for textual, symbolic information only, but I strongly doubt that nonsymbolic information (such as learning that jump) is attained at just one bit per second, or we would not have the saying that a picture is worth a thousand words.--Jasper Deng (talk) 06:28, 4 June 2017 (UTC)[reply]

Is sweetness intensity associated with sucrose quantity?

[edit]

Is sweetness intensity associated with sucrose quantity? Is there a sucrose quantity upper threshold at which the sweetness "tastes the same"? Or is sweetness perception variable from individual to individual? Is there a point at which sweetness becomes unbearable, or is that subjective? All I know is, I've never liked the icing and fillings of bakery goods. They are too sweet for me. It's one reason why my favorite type of doughnut is the plain glazed doughnut, eaten with water to dilute the sweetness. I think this low-sweetness-tolerance threshold is genetic, because my father and mother also prefer mildly sweet things so icings, fillings, jams and fruit preserves, fruit juices, fruit smoothies are rarely consumed in the house. 50.4.236.254 (talk) 23:13, 31 May 2017 (UTC)[reply]

The Wikipedia article with the surprising title of sweetness gives the common scale by which sweetness is measured, with pure sucrose given a sweetness value of "1.00" Concentrations of sucrose less than pure will, of course, have values less than one. Here are some external sources which go into more detail on methods for determining sweetness. --Jayron32 23:20, 31 May 2017 (UTC)[reply]
It's difficult to separate genetic and environmental factors in such cases. You may have inherited genes making you more sensitive to sweetness, or you may have become habituated to a lower level of sweetness because you were raised in a family culture, and retain those behaviours, exposing you to lower levels of sugar than the average, or both may be true. ObPersonal, but when in the past I cut down the amount of (cane) sugar I consumed in, for example, tea, at first it tasted insufficiently sweet, but after a few weeks became the new normal, and tea with the previous amount of sugar tasted unpleasantly sweet. I have no doubt that if I were to reverse that change, the reverse would occur. {The poster formerly known as 87.81.230.195} 94.9.8.213 (talk) 23:37, 31 May 2017 (UTC)[reply]
  • The most important reason this doesn't work is that perceived sweetness depends very heavily on context. If a given amount of sugar is mixed with a sour of bitter substance it will be perceived as less sweet. For example if you took the lemon out of lemonade, the amount of sugar in it would taste icky sweet to most people. Looie496 (talk) 15:06, 2 June 2017 (UTC)[reply]