GHC

Joined 17 February 2020
5,356 bytes added ,  4 April 2020
no edit summary
No edit summary
No edit summary
Line 1: Line 1:
... the road towards learning can be seen as a bifurcation. This occurs at the thing you're trying to learn (show an image of a road leading to the topic and bifurcating depending on the type of topic) You can look at the brain for queues about ...
== ON MEMORY ==


No one has to read about how cognitive scientists think the brain processes information (absorbing, storing, processing and recalling) and then complement that theory with studies about recall, error minimization or learning techniques done in learning theory in order to start learning. This would be a recursive nightmare - where does learning start if you're required to be able to learn in the first place?! No. It comes naturally doesn't it? You might remember when you were fresh out of the womb and couldn't walk from there to anywhere - or should I say, you might remember the idea of it, but you actually, factually speaking and theoretically sounding, can't really remember jack shit about it. Let me start here then. We don't really know why it is so, and this is the general rule in regards to brain mechanics - we don't know that much about it. But anyhow,  it has to do with neural networks and/or the absence of these and their meaningful connections. Basically, you had no saved games yet. But! We know about what comes straight after that. In your (roughly) first three turns around the sun you will have learned the basics of language, having solidified your mother tongue in a way that you will never manage to do with any other language from then on. This is in part because you won't have the same processing structure available. You are born with an unfair boost of neurons that dissipate with age which cognitive scientists believe gives you the superpower of learning a language without any prior knowledge. Every other language will be learned by analogy to your first language. You will have learned how to move as well. Your motor cortex, which is a slice of neurons that's roughly in the middle of your brain will have developed so many connections through trial and error, stumbling and falling so many times that you will have brute-forced yourself into standing. Aided by specified cortexes and lobes and circuitry of neurons with enough cable line to go around the world about 4.4 times (earth's circumference is more or less 40,000 km and an average 20-year-old man has 176,000 km of myelinated fibres - this is the white matter of the brain, or rather, the cabling that connects the grey matter which holds the compute power). You will have learned to identify people by their face, voice, movement and with some margin of error, just from looking at their backs. You will be able to output smiles, pouts, crying and all sorts of complex human emotional expression. What I am trying to say is that you learn all the time independently of you wanting to do so or not. That's it. You can go. Hmm. Well, maybe before you blame this article for your next failed exam I should be more specific... Ok, wait! Although your brain catches and learns patterns automatically these patterns are somewhat specific - like those play-doh factory toys, the ones that make the doh into a specific shape, squares, circles, triangles, etc. Your brain is shaping the signals coming in into meaningful information which is dependent on your previous recorded experiences. Meaningful in regards to what it can compute. Your brain has all these specific circuitry which can process and store information in a specific way. Thanks to the gazillions of years that the brain has been rolling back and forth in the ocean of evolution and randomness, it can do very interesting things indeed. Maybe you might have to run from shit that is running after you or you might have to be the one doing the chasing. Evolution is a dumb impartial algorithm. It selects for whatever machine seems to survive and propagate its genes better - and if you let the simulation run for long enough even marginal advantages make huge differences. This is to say that we can't forget that our brain evolved or was optimized for a specific environment type. This is important to acknowledge because the reason why you're reading this article is that you might be trying to optimize your learning, and you're probably not trying to learn how to climb a tree, or detect tigers better. You're probably trying to learn math, or physics or memorizing something that you need to know for an exam.
No one has to read about how cognitive scientists think the brain processes information (absorbing, storing, processing and recalling) and then complement that theory with studies about recall, error minimization or learning techniques done in learning theory in order to start learning. This would be a recursive nightmare - where does learning start if you're required to be able to learn in the first place?! No. It comes naturally doesn't it? But here you are.


When you're learning, you're making connections between the newly acquired material and other information already in your memory. These connections help you because they make new knowledge findable later on.
You might remember when you were fresh out of the womb and couldn't walk from there to anywhere... Actually, you might remember the idea of it. But you actually, factually speaking and theoretically sounding, can't really remember jack shit about that actual time. Let me start here then - with the memories you think you remember. We don't really know why this is so, and this will be the general rule in regards to brain mechanics - we don't know that much about it. But anyhow, it has to do with neural networks and/or the absence of these and their meaningful connections back when you were little.  
Rather, these connections serve as retrieval paths: The theory is that whenever you try to retrieve a specific piece of memory your brain follows crumbs or pieces of thread that lead to this particular memory (some sort of threshold is probably used as that is how neurons work - connections get aggregated and strengthen the closer they are by some negligible value or biological equivalent. So can there be different threads to the memory in question? Can some threads be better/faster than others? Context-dependent learning: basically the data shows data learning is context-dependent and aided by context as well - studies using randomized control groups show that mimicking the learning context during testing aids retrieval for the people with whom the learning context matches the testing context (Eich, 1980; O, 1985), Godden and Baddeley (1975), Grant et al., 1998; Balch, Bowman, & Mohler, 1992; Cann & Ross, 1989, Schab, 1990; Smith, 1985; Smith & Vela, 2001. A Smith (1979) should in his studies thought, what matters is not the physical context per se but the psychological context - he urged participants to remember and picture the learning-context while doing that test task in a different place and when matching their results to a group that did the test in the same learning-context, the results were equivalent. One could say that any context is simply psychological as cues and signals are unravelled and built into the mind's picture that we all "see"This means that you can get the benefits of context-dependent learning through a strategy of context reinstatement - a strategy of re-creating the thoughts and feelings of the learning episode even if, at the time of recall, you're in a very different place. What matters for memory retrieval is the mental context, not the physical environment itself to the extent that you can abstract and extract yourself from the latter.


Pages 224
But!
 
We know about what comes straight after that. In your (roughly) first three turns around the sun you will have learned the basics of language, having solidified your mother tongue in a way that you will never manage to do with any other language from then on. This is in part because you won't have the same processing structure available. You are born with an unfair boost of neurons that dissipate with age which cognitive scientists believe gives you the superpower of learning a language without any prior knowledge (among other things). Every other language will be learned by analogy to your first language.
You will have learned how to move. Your motor cortex, which is a slice of neurons that's roughly in the middle of your brain, will have developed so many connections through trial and error, stumbling and falling so many times that you will have brute-forced yourself into standing. Aided by specified cortexes and lobes and circuitry of neurons with enough cable line to go around the world about 4.4 times (earth's circumference is more or less 40,000 km and an average 20-year-old man has 176,000 km of myelinated fibres - this is the white matter of the brain, or rather, the cabling that connects the grey matter which holds the computing power).
You will have learned to identify people by face, voice, movement and with some margin of error, just from looking at their backs.
You will be able to output smiles, pouts, crying and all sorts of complex human emotional expression.
 
What I am trying to say is that you learn all the time independently of you wanting to do so or not. That's it. You can go.
Hmm.
Well, maybe before you blame this article for your next failed exam I should be more specific... Ok, wait!
Although your brain catches and learns patterns automatically these patterns are somewhat specific - like those play-doh factory toys, the ones that make the doh into a specific shape, squares, circles, triangles, etc. Your brain is shaping the signals coming in into meaningful information which is dependent on your previous recorded experiences as well as your actual hardware - your brain. Meaningful in regards to what your brain can compute. Your brain has all these specific circuitry which can process and store information in a certain way with some degree of plasticity (some parts can take the function of other parts in case the latter are damaged). Thanks to the gazillions of years that the brain has been rolling back and forth in the ocean of evolution and randomness, it can do very interesting things indeed. Maybe you might have to run from shit that is running after you or you might have to be the one doing the chasing. But don't forget that evolution is a dumb impartial algorithm. It selects for whatever machine seems to survive and propagate its genes better in the current environment. And if you let the simulation run for long enough even marginal advantages make huge differences. This is to say that we can't forget that our brain evolved or was optimized for a specific environment type. This is important to acknowledge because the reason why you're reading this article is, I should bet, that you are trying to optimize your learning. And you're probably not trying to learn how to climb a tree or find the most optimal way to juke a tiger. You're probably trying to learn math, or physics or memorizing something that you want to know.
 
Let me then share with you some of the things cognitive psychology found out from modelling and working closely with psychology and neuroscience.
 
When you're learning, you're making connections between the newly acquired material and other information already in your memory. The first 'rule' or study hint is that new memories that get conjugated or mashed with older memories will be easier to retrieve. A strategy that uses this fact is called the method of loci (this method is a mnemonic device adopted in ancient Roman and Greek rhetorical treatises in the anonymous Rhetorica ad Herennium, Cicero's De Oratore, and Quintilian's Institutio Oratoria. From Wikipedia <3). In the method of loci one simply associates the new information by means of a narrative, a story, to a physical place that is already well consolidated. So, for example, imagine you have to quickly remember a string of objects. The way you would use this technique would be by associating each object or every two objects or so with each room of your house by telling a story that puts the object in the room. If the order matters, you can even add a number to each (room,object) pair. These connections help you because they make new knowledge findable later on by association to the already present memories.
 
Older memories serve as retrieval paths. The theory is that whenever you try to retrieve a specific piece of memory your brain follows crumbs or pieces of thread that lead to this particular memory (some sort of threshold is probably used as that is how neurons work - connections get aggregated and strengthen the closer they are by some negligible value or biological equivalent). So can there be different threads to the memory in question? Can some threads be better/faster than others? Is there such a thing as context-dependent learning? Yes. Data shows that learning is context-dependent and aided by said context as well - studies using randomized control groups show that mimicking the learning context during testing aids retrieval for the people with whom the learning context matches the testing context (Eich, 1980; O, 1985), Godden and Baddeley (1975), Grant et al., 1998; Balch, Bowman, & Mohler, 1992; Cann & Ross, 1989, Schab, 1990; Smith, 1985; Smith & Vela, 2001. A Smith (1979). According to some of these studies what matters is not the physical context per se but the psychological context. In a specific study, participants were to remember and picture the learning-context while doing a test task in a different place and when matching their results to a group that did the test in the same learning-context, the results were equivalent. One could say that any context is simply psychological as cues and signals are unravelled and built into the mind's picture that we all "see".  This means that you can get the benefits of context-dependent learning through a strategy of context reinstatement - a strategy of re-creating the thoughts and feelings of the learning episode even if, at the time of recall, you're in a very different place. What matters for memory retrieval is the mental context, not the physical environment itself to the extent that you can abstract and extract yourself from the latter.
 
Other experiments have shown that memory is context-dependent in terms of the material itself. On being given a specific sentence that participants were asked explicitly to remember, participants were shown to remember the sentence more accurately when given a cue that was synonymous or indicative of the context of the sentence. For example, 'the man lifted the piano' was easier to recall when given the cue 'something heavy'. While if given the sentence 'the man tuned the piano', it was easier to recall if given the sentence 'nice sound' (Barclay, Bransford, Franks, McCarrell, & Nitsch, 1974). The cue was also shown to only be effective when it was in line with what was being stored in memory. This pattern is often called 'encoding specificity' (Tulving, 1983). So the stimulus is wrapped or encoded together with its physical context as well as its semantic context.
 
Another keyword in these scientists lang is 'spreading activation'. Where suppositions about brain mechanics are made, which of course, follow from experiment. A conceptual node that represents a memory object is taken to exist in a large network of other nodes. These nodes are activated upon receiving strong input signals. Once activated, the node can in its turn activate other nodes by spreading energy via its associations to its sibling nodes. If this energy reaches a nodes response threshold then the node fires away as well, causing further spreading activation. Activation is thought to accumulate, so even if a node doesn't fire per se it can still add, or summate, to other nodes and further propagate the activation. This node firing is important because scientists think that this is what brings attention to the node itself, or rather, how memory cascades into the conscious. The theory posed in the previous paragraphs can now be understood from the perspective of this spreading activation in a network of nodes. Retrieval cues and context reinstatement will aid memory recall because these involve firing close-enough nodes to the actual memory that eventually the target memory is bound to fire as well.
 
Another cool trick that a cognitive scientist could tell you about is semantic priming (and other sorts of priming). This technique involves priming yourself or reading cues before a task (i.e. exam), which will result in a faster response or recall during said task. This priming can be semantic - like the piano example, where you think of semantically related or similar-in-context themes to the thing you're trying to recall. It can also be repetition priming, where you read and re-read a sentence eventually creating priming links between words, one causing activation of the other, thus conducing memorization.
 
There is an inconvenient truth that is an outcome of the memory system that we humans are running (which is basically intuitionistic and not rationalistic at all - with the causality chain being hardcoded in the form of mechanical bumps of energy and chemicals - in other words, there is no thinking machine inside the thinking machine) involving some brain parts like the temporal lobes, the amygdala, the hippocampus and the rhinal cortex among others. Implicits memories and the 'Illusion of truth' are one inconvenient fact. We tend to remember things by context and not by detail. So spreading activation often induces us in error or close-enough situations especially when it comes to remembering facts, similar objects or grainy details like numbers, equations or names. These are often bundled up together anyway so it's easy to get them confused. One possible solution would be to associate them with very different things. Personally I've heard of very good math professors (including Judea Pearl who once spoke about his favourite math teacher in his childhood who influenced him greatly) who taught their material chronologically alongside the story of their inventors and historical context of the invention. This seemed to have the wanted effect - now the numbers and equations are not just numbers and equations but their solidified and differentiated by the aid of their individual history.
 
== ON PROBLEM SOLVING TECHNIQUES ==
474
474
Anonymous user