flat assembler
Message board for the users of flat assembler.

Index > Heap > how you learn, getting the know knowledge

Goto page 1, 2  Next
Author
Thread Post new topic Reply to topic
sleepsleep



Joined: 05 Oct 2006
Posts: 8864
Location: ˛                             ⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣Posts: 334455
sleepsleep
i suspect, we just fail to analyze how we actually learn.
there are lot of knowledge out there, one thing for sure is, each different kind of knowledge is best use with different kind of method to understand and using it.

is the people who best learn are the one who possess the best memory to memorized all sort of information?

if the idea is to understand, then to apply knowledge. i just felt we totally lack the critical information in this area.
Post 11 Feb 2009, 20:21
View user's profile Send private message Reply with quote
Borsuc



Joined: 29 Dec 2005
Posts: 2466
Location: Bucharest, Romania
Borsuc
Cool stuff. That thing doesn't make a consciousness but it makes an unconscious/instinctual part well enough Cool

_________________
Previously known as The_Grey_Beast
Post 12 Feb 2009, 01:10
View user's profile Send private message Reply with quote
revolution
When all else fails, read the source


Joined: 24 Aug 2004
Posts: 17247
Location: In your JS exploiting you and your system
revolution
Experience is the best teacher. Get out there and experience things, try things, do things.
Post 12 Feb 2009, 01:12
View user's profile Send private message Visit poster's website Reply with quote
bitRAKE



Joined: 21 Jul 2003
Posts: 2887
Location: [RSP+8*5]
bitRAKE
Most people are good kindergarteners - show them anything and they can copy it - rote learning.

Experience is what you get when you're trying for something else. Laughing
Post 12 Feb 2009, 05:39
View user's profile Send private message Visit poster's website Reply with quote
DJ Mauretto



Joined: 14 Mar 2007
Posts: 464
Location: Rome,Italy
DJ Mauretto
i usually memorized all sort of information,is a natural way for me
i had a good memory since when was young My mother often sent me to do the supermarket shopping with long lists, I am ashamed to show
my list so 'learned by heart the entire shopping list, now I am storing natural data,
also important is the visual memory,
I have always been, I think that everything depends
from experience that you had early in life Wink

_________________
Nil Volentibus Arduum Razz
Post 12 Feb 2009, 09:32
View user's profile Send private message Reply with quote
Borsuc



Joined: 29 Dec 2005
Posts: 2466
Location: Bucharest, Romania
Borsuc
oh I thought this thread was about how the process in learning works Razz (ie: make some 'learning' simulator or program that can simulate this)

_________________
Previously known as The_Grey_Beast
Post 12 Feb 2009, 17:38
View user's profile Send private message Reply with quote
bitRAKE



Joined: 21 Jul 2003
Posts: 2887
Location: [RSP+8*5]
bitRAKE
Douglas B. Lenat's Automated Mathematician and Eurisko had me intrigued from a young age about software that learned.

Try this: What does Cyc know?
Post 13 Feb 2009, 01:58
View user's profile Send private message Visit poster's website Reply with quote
sleepsleep



Joined: 05 Oct 2006
Posts: 8864
Location: ˛                             ⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣Posts: 334455
sleepsleep
some of text i wrote to extend my knowledge regarding how i learn

Learn by desire
extend the know knowledge by desire to know more.
eg. when you see the word "important", you might feel like you want to know another word that have the same meaning, so you google around or checking dictionary, then you find the work critical. So something new is added into your knowledge pool.

Learn by remember
remember is part of learning, if you couldn't recall something which you want to use or you did know, it is resemble as don't know.

Learn by experience
To test is part of experience, eg. not all dogs tame, so, maybe in experience, you see some tame dog before, but in another moment of life, you saw an erratic dog, so by using memory plus experience, you learn that all dog got different behavious, and by remember their face then we judge a dog by using our memory whether it is tame or not.

Learn by forming concept
Forming a concept means to sum up some part of our already know knowledge and form a category or a special new concept, conclusion, idea based on it. If 5 + 3 = 8, 4 + 4 = 8, 2 * 4 = 8, 16 / 2 = 8, then we got a concept that, the output of above sum is equal.

Learn by accepting rules
By remember the rules and accept it, we learn something. Eg. in HTML language, the first thing we see is probably <HTML> tag. There are lot of tag, later we know that in order to bold a text, we need to put a start tag <b> and our text in between, then a close tag </b>, so text in between will get bold. But why a slash in end tag, because an end tag must have a slash, so this is part of the rule that we required to follow in order to use it.

Learn by guessing and believe
Eg, a person put a dice under 3 cups, then switch / rotate them several times, then ask you to choose, which cup contains dice? So, what happen here is we guess, base on what we see, then by believing it, we tell him, the dice is under cup number 2.

please give me more input on how you learn. thank you.
Post 13 Feb 2009, 11:57
View user's profile Send private message Reply with quote
Madis731



Joined: 25 Sep 2003
Posts: 2141
Location: Estonia
Madis731
Cyc is actually a relational database. It can't read Wikipedia and learn. AI should be able to do at least that.
i.e.
1) User asks what Fe is
- AI doesn't have it in his huge database of relational data
2) AI 'goes browsing' en.wikipedia.org\wiki\Fe and reads all data
3) What makes AI special is not Wiki-query, but the ability to extract this "Iron chemical element with the symbol Fe" and change it to:
4) AI replys: "Fe is iron and its a chemical element"

Other practical uses include "paste a loop template to clipboard with FASM syntax"; "try all possible typing-mistakes of 'Tallinn' on Google"; etc.
Post 13 Feb 2009, 12:16
View user's profile Send private message Visit poster's website Yahoo Messenger MSN Messenger Reply with quote
tom tobias



Joined: 09 Sep 2003
Posts: 1320
Location: usa
tom tobias
The first step is to recognize that "learning" implies recognition by someone, oneself, or another individual or group, that a particular, defined, delimited response, i.e. motor behaviour, occurs more rapidly (i.e. more efficiently) with subsequent presentations of a unique stimulus (e.g. cue) as compared with the initial presentation of this identical stimulus. Consider learning Chinese characters. Presentation of particular character, on a flash card, for example, with a requirement to commence phonating the sound in Mandarin Chinese, and explain the meaning in English (the "motor behaviour", above) requires, let us imagine, two minutes, initially, and subsequently, two seconds, then, ultimately, on attaining literacy, two hundred milliseconds. We call this improvement, this decrease in time required to initiate the expected motor skill, "learning".
Then, we may reformulate sleep^2' question this way: what is the cellular mechanism responsible for this observable, measurable decrease in time needed to commence execution of a sterotyped response? The synaptic transmission time between neurons has not changed. The axonal conduction velocity along the length of the neuron is unchanged. The neuronal geometry is unchanged. The neuronal pathways linking the sensory gateways to the motor outflow tracts have not been altered. So, then, if none of those parameters have changed, what is the cellular mechanism responsible for the improved efficiency, i.e. the decreased time needed to commence a response?

To address this question, one needs to employ a system more amenable to scrutiny than the human brain, in particular, one needs to work with neurons in invertebrates, ideally, or "simple" vertebrates, i.e. in animals with far fewer neurons than most vertebrates possess. One such model is the goldfish, which possesses large "Mauthner" cells, neurons specialized for rapid escape, following reception of an auditory stimulus. Here's a reference for those intrigued by such inquiries, and I would suggest entering "cellular basis learning" into google for further information.
goldfish

In my opinion, one can envision a submarine, with a crew of thirty, and consider the very first time the crew encounters, in a drill, perhaps, the need to exit the deck, enter the hatch, and descend to safety, as the submarine prepares to make an emergency dive below water. On the second and third encounters of this sort, the crew will have recognized the need to maintain a minimum quantity of sailors on the deck, when the ship surfaces, so as to facilitate rapid deployment below the hatch door, upon hearing the alarm klaxon, in order to successfully evade the rapidly encroaching sea water. In other words, I envision a cellular process whereby multiple neurons, slowly conducting with unmyelinated axons, are bypassed from the normal circuitry, and a direct linkage created between the alarm input, and the output cell--> for the goldfish, this would be the Mauthner cell.
With regard to computer software's ability to "learn", it has been a goal, at least since the late 1940's, when Arthur C. Clarke published a collection of short stories, including the gist of 2001. One could write volumes on the subject, and still not produce a program capable of "learning".
Post 13 Feb 2009, 12:44
View user's profile Send private message Reply with quote
Madis731



Joined: 25 Sep 2003
Posts: 2141
Location: Estonia
Madis731
Sorry, but a goldfish isn't capable of learning Japanese: http://www.jstage.jst.go.jp/article/biophys/41/2/80/_pdf
Isn't there a more well-known language this article was translated?

FASM AI FTW Smile

Fai-fai Very Happy "FAI" Razz


EDIT: On a more serious side of Tom Tobias's text, I would like to think of an AI-in-a-box, where its only IO is console so it doesn't have the six senses we do. The "submarine-crew" experiment can be conducted in our heads and we can think of multiple options. One of them being to enlarge the hatch to fit them in more quickly, the other one of course having less of them there in the first place Smile.
What AI needs to know about this situation is overwhelming:
1) Humans try to preserve their life
2) Humans can't stay under water for extended periods
3) The crew needs to be in the submarine for the sub to be operable
4) Submarine is valuable to humans
5) Danger means one needs to start worrying about his/her wellbeing
6) Submarine is safer than water
7) Other humans need to enter through the same hatch - the priorities?
Cool ... (the list goes on)
There are other things what one doesn't generally think about
*) Humans can move
*) Humans have a physical location
*) Location is on the sub
*) Movement toward the hatch - the shortest/quickest/safest route...
*) Humans can open/close the hatch
*) You need to open the hatch to enter the sub
*) Closing the hatch means preserving ones life
**) OTOH closing yourself in a 'can' might be scary at first (is AI scared?)
...
*) You need recognize alarm, tie it to problems
*) ...
Post 13 Feb 2009, 13:19
View user's profile Send private message Visit poster's website Yahoo Messenger MSN Messenger Reply with quote
bitRAKE



Joined: 21 Jul 2003
Posts: 2887
Location: [RSP+8*5]
bitRAKE
Douglas' premise (or pre-miss if you wish Laughing ) with Cyc was to produce a base of rules to feed into a discovery engine. This base was suppose to be the "common sense" of humanity that would allow the discovery engine to process a set of encyclopedias.
Post 13 Feb 2009, 16:12
View user's profile Send private message Visit poster's website Reply with quote
sleepsleep



Joined: 05 Oct 2006
Posts: 8864
Location: ˛                             ⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣Posts: 334455
sleepsleep
Quote:

So, then, if none of those parameters have changed, what is the cellular mechanism responsible for the improved efficiency, i.e. the decreased time needed to commence a response?

i guess brain memory plays a big part here (imo) i am not scienties or watsoever.

but our memory could tricks us too quite easily. let me show you how.
eg. covered up an empty bottle and ask your friend to fill in tap water into it. later ask him/her to put it in on shelf. now get him to do something else, then you secretly pour out all the water inside that bottle and put it back. then ask your friend to take the bottle and put it on another shelf.

wat happened here is, your friend would use exactly da strength he remember when he put that bottle on shelf.


another part of learning is, we could easily see students who poor in academic but do it rather well in gaming. i believe education is suppose to find wat strength inside each student and enchance them rather than forcing them to follow one kind of rulez.

if he did well in starcraft, monopoly, transport tycoon, it means, the game rules appeal to his logic, so by bringing and bridging the game rules into real life situation, he/she would excel without questions, imo.

everybody can be programmer, can code an OS, it just they need the platform to support their rules, instead of forcing them to learn the rules. technology should play assisting role instead of decisive role. imo too.
Post 14 Feb 2009, 19:12
View user's profile Send private message Reply with quote
sleepsleep



Joined: 05 Oct 2006
Posts: 8864
Location: ˛                             ⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣Posts: 334455
sleepsleep
cool, i just thought of a new concept.

imagine if we could apply the transport tycoon game concept on programming.

http://en.wikipedia.org/wiki/Transport_Tycoon
http://www.openttd.org/en/

in the world of transport tycoon, we got 4 basic means of transport mechanism.

air (planes), water (ship/tank), road (bus/truck) and trains.

so, we might call a 4 coaches train as EAX.
and we got 8 trains here.

each coach can be marked as to load with passenger or goods.
passenger is unsigned and goods is signed.
so, maximum number that we can load into coah passenger is 255.
for goods coah, we can reserve place for goods, which will be unsigned number then.

int is the route for train to go. for any time, game only allow one train to run.

for example,
mov ax,5
in game GUI mode, means, user need to select train EAX, place a plate 5 on railroad, so when the train passes through the railroad, it would load 5.

station could be identified as function. when user want to perform certain operation, he can drive the train into particular station.

.... something like that i guess. more interesting. so programming will get more like playing game. and parallel programming will be ease...

if we are more to object and visual aware, why the hell we use word/character to program?
Post 15 Feb 2009, 04:31
View user's profile Send private message Reply with quote
bitRAKE



Joined: 21 Jul 2003
Posts: 2887
Location: [RSP+8*5]
bitRAKE
sleepsleep wrote:
if we are more to object and visual aware, why the hell we use word/character to program?
Try working in more than three dimensions visually - on a 2D screen no less. Perspective can make some things easier and other things harder. I took some notes on a visual system for SIMD instructions, but attempting to convert some existing code convinced me of the futility in the effort. The graphical system was too restrictive/rigid for the range of instruction use.

A moderately experienced programmer can find novel uses for instructions beyond the explicit intent. Using multiply to rearrange bit-fields within a single instruction (SWAR), for example. Or stacking flags from multiple tests for use on a single branch.

_________________
¯\(°_o)/¯ unlicense.org
Post 15 Feb 2009, 06:49
View user's profile Send private message Visit poster's website Reply with quote
Borsuc



Joined: 29 Dec 2005
Posts: 2466
Location: Bucharest, Romania
Borsuc
sleepsleep wrote:
.... something like that i guess. more interesting. so programming will get more like playing game. and parallel programming will be ease...
That's funny but it would take a buttload of time.
Remember sleepsleep, it is not the writing process that is important in programming, but the THINKING. No matter if you use english 'mov' instructions (ok, not really english Laughing) or a 'train' to 'see' the instructions, you'll still need that phase called thinking before-hand. Wink

_________________
Previously known as The_Grey_Beast
Post 15 Feb 2009, 16:27
View user's profile Send private message Reply with quote
Madis731



Joined: 25 Sep 2003
Posts: 2141
Location: Estonia
Madis731
We drifted a bit off-topic, but to get us back on track, I will comment a bit more the first post. I will mix here some of our brain and some of AI Smile

First I would say that the best memory is not the primary objective to achieve. You need to be able to GENERALIZE, which means extract a general action or object out of a detailed one (i.e. '98 Audi A6 is a car) and you also need to be able to do the opposite: extract details out of general objects.
The next most important part of learning is comparing these general objects/events to others of the same type.
Easiest to give an (AI) example here:
I ask AI: "What colour is the Sun?" and "Is Earth smaller than the Sun?"
Sun->colour is a simple database lookup, (which fails if you ask the colour of love Smile) the answer would probably be "Yellow" or something similar.
Smaller is a comparator and therefore AI first needs to see if the Earth and Sun are a) compatible and b) if size property can be asked. If they both hold true, the simple answer would be "Yes".

We (and therefore AI) must be able to compare and generalize to be able to think. Of course its not the absolute truth.
Post 15 Feb 2009, 16:38
View user's profile Send private message Visit poster's website Yahoo Messenger MSN Messenger Reply with quote
sleepsleep



Joined: 05 Oct 2006
Posts: 8864
Location: ˛                             ⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣Posts: 334455
sleepsleep
Quote:

Remember sleepsleep, it is not the writing process that is important in programming, but the THINKING. No matter if you use english 'mov' instructions (ok, not really english ) or a 'train' to 'see' the instructions, you'll still need that phase called thinking before-hand.

yeah, i agree that thinking is extremely important in programming, but i am focusing on after you got the idea/thoughts and you want to realize/present/output it into actual product.

to use a programming language, may it c / c++ / pascal / asm or etc, it means, he needs to study how to use it, all the main syntax and so on. if doing it through game, i for sure believe, it would be totally different and enjoyable.
Post 15 Feb 2009, 16:47
View user's profile Send private message Reply with quote
LocoDelAssembly
Your code has a bug


Joined: 06 May 2005
Posts: 4633
Location: Argentina
LocoDelAssembly
Not sure if it is useful but here it goes: Course | Machine Learning (Stanford)
Post 15 Feb 2009, 17:28
View user's profile Send private message Reply with quote
tom tobias



Joined: 09 Sep 2003
Posts: 1320
Location: usa
tom tobias
sleep^2 wrote:
...is the people who best learn are the one who possess the best memory ...
...
i agree that thinking is extremely important in programming, but i am focusing on after you got the idea/thoughts and you want to realize/present/output it into actual product.

One may wish to be a tiny bit careful when using these three words:
learning
memory
thinking
Yes, I agree, to a certain extent, this is a kind of fussiness on my part. More of an English discussion than a philosophical discussion I suppose. sorry.
Point is this: These three words convey three very different physiological processes, regardless of whether or not they possess some overlap linguistically, or in the context of computer science.
thinking: act of negating that which is immediately before us. aka cogitation, reflection, contemplation.
All of these words imply a neuronal operation performing something analogous to a British Museum Search algorithm.
Memory: easiest of the three to comprehend: just like RAM or cache.
learning: the most difficult concept to describe: this operation reminds one of human fraility. We learn foreign languages as youngsters, whose brains have not yet fully matured. By age 12 or 14 or so, the brain is fully developed, and the human ability to master foreign languages is greatly hindered. This suggests, to me at least, that "learning" has something to do with fundamental neuronal circuitry, rather than "plasticity"--i.e. change in efficiency of individual or group synaptic connections.
In other words, in my opinion, not a fact, "learning" in adults, and children are not synonymous, from a cellular physiological perspective. I have no idea how to define "learning" in a computer program. Can an inanimate object, without consciousness, learn? This topic was a favorite theme of Science Fiction authors for the past six decades....
Post 15 Feb 2009, 20:36
View user's profile Send private message Reply with quote
Display posts from previous:
Post new topic Reply to topic

Jump to:  
Goto page 1, 2  Next

< Last Thread | Next Thread >
Forum Rules:
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You can attach files in this forum
You can download files in this forum


Copyright © 1999-2020, Tomasz Grysztar.

Powered by rwasa.