Probably because you suggested that someone who has their own stuff to do go get a college degree to resolve one problem they think they see in education. Also, your second sentence sounds a bit like an advertisement for Charles Darwin Universities Bachelor of Teaching and Learning Preservice.
To do better, just say “I’m studying education, and they’re spending a lot of time on this stuff called ‘deep learning,’” and then maybe you could spend a few sentences actually talking about what you’re studying and explaining it to Robin, rather than expecting him to sign up for classes to find out.
So it’s a “9 point” positive to say something that reiterates a commonly perceived problem and offers no solution and also makes a factual error, but to direct someone to a place that actually addresses their problem is a −1. Cool, I’m getting a feel for the website now, cheers.
Let’s try this.
It’s actually pretty easy to test for deeper learning. For example multiple choice questions have previously been considered as examples of shallow learning, or if you prefer shallow testing, and in the past that was accurate and indeed in some still existing multiple choice questions there isn’t a path towards deep learning. However consider this question.
The 11 letters in the word PROBABILITY are written on 11 pieces of paper, and a piece of paper chosen at random from a bag. Which of the following statements are true?
a)The probability of selecting a “B” is less than the probability of selecting an “I”.
b)There is a greater chance of obtaining a consonant than of obtaining a vowel.
c)A vowel is less likely than a consonant.
d)If you repeated the experiment a very large number of times, approximately 63% of the results would be consonants.
Make your selection, note that you may select more than zero answers.
Now since “advertising” for CDU might be deemed as somewhat negative, am I permitted to share the details of a book one could read to understand how deep learning for mathematics can be taught to middle school students? Or would that be advertising also?
Elementary & Middle School Mathematics : Teaching Developmentally by John A. Van De Walle, Karen S. Karp and Jennifer M. Bay-Williams, published by Pearson International.
So it’s a “9 point” positive to say something that reiterates a commonly perceived problem and offers no solution and also makes a factual error,
If you’re Robin Hanson, sure. Also note that the comment is from ’07: the voting system wasn’t put in place until after the move from Overcoming Bias (Hanson’s blog) to LessWrong. If a similar comment was made today it would probably be voted up much higher. Or maybe downvoted, who knows.
but to direct someone to a place that actually addresses their problem is a −1.
Here, I gave you an upvote. Now it’s a 0. Karma means too little to stress out over like this.
Cool, I’m getting a feel for the website now, cheers.
Telling someone to read a book is still probably expecting too much of them. An online article is pretty much the limit, and it’s even better if you can tell the other person what you want them to learn in your own words. Referencing a book is good, though.
Cheers Manfred, Hopefully I’ve given a glimpse of how one can apply and test for deep learning… and if I a “lowly” first year undergrad in education can do it, seems it’s probably not that hard. [smirk]
I’m getting a few mixed signals from the site, and that’s to be expected.
Some want high detail, others want little. Some will downgrade me for restating something that’s been said, others don’t seem to be downgraded for saying something that’s common knowledge. I’ll get the hang of it. I’m well aware that joining an internet site/forum/community involves some amount of fitting in, some amount of taking hits cause you’re the new guy or just aren’t getting it, and some amount of adjusting to the norms, indeed some good old fashion relationship building too.
I was hoping this might be a place where solutions are discussed more than problems are complained about. Perhaps that’s true and I just haven’t discovered it yet, don’t know.
I remain open minded and am working my way through the sequences, they’re good reading.
Disconfirmation bias—love that concept and discovered it on lesswrong, hadn’t found that one in the Social Psych or Education Psych textbook yet. Even if I get too peeved off and leave to never return, LW will be remembered fondly for that one concept alone.
Could you give an example of “shallow learning” alongside one of “deep learning,” and explain the difference? “Deep learning” definitely sounds like something that’s better than “shallow learning,” but you haven’t made if very clear what it actually is.
Desrtopa, sure thing mate.
Deep learning example :
The 11 letters in the word PROBABILITY are written on 11 pieces of paper, and a piece of paper chosen at random from a bag. Which of the following statements are true?
a)The probability of selecting a “B” is less than the probability of selecting an “I”.
b)There is a greater chance of obtaining a consonant than of obtaining a vowel.
c)A vowel is less likely than a consonant.
d)If you repeated the experiment a very large number of times, approximately 63% of the results would be consonants.
Make your selection, note that you may select more than zero answers.
Shallow learning example.
The 11 letters of FOUNDATIONS are written on 11 pieces of paper, and a piece of paper chosen at random from a bag. What is the probability that an “O” is selected?
a) 3⁄12.
b) 1⁄11.
c) 2⁄11.
d) 2⁄12.
Select only 1 answer.
The Deep learning example uses the word “probability” that’s a way to prime the student to thinking in terms of probability, it is an interconnection, it enhances learning, the word “foundations” doesn’t do this. The question is “which of the following statements are true?” – that’s a question that is more open than “What is the probability that an “O” is selected?”—open questions evoke deep learning better than closed questions. The answer selections in the deep learning are worded, they require interpretation, they need an understanding of consonants and vowels – which again provides an interconnection with English, and interconnections are deep learning. Whilst the shallow learning has 4 numbers for options, they require no interpretation and they don’t interconnect with English as much as does the deep learning example. The Deep learning questions “make you selection, note that you may select more than zero answers” gives the reader a pause… how many can I select, what does more than zero mean, it requires some interpretation, could be 1, 2, 3, or 4! The “Select only 1 answer” doesn’t need interpretation, it’s closed – just 1. Now about the answers themselves. Shallow learning multiple choice questions typically have 2 options that are readily visible as incorrect and can be quickly discarded, 3⁄12 can be quickly discarded because there aren’t 3 O’s nor are there 12 letters. 1⁄11 can be quickly discarded because there are 2 “O”s. 2⁄11 is the correct answer and so the person doesn’t even need to assess answer d. Where as in the deep learning each answer needs to be assessed for their truth value, and no answer can be quickly discarded.
Probably because you suggested that someone who has their own stuff to do go get a college degree to resolve one problem they think they see in education. Also, your second sentence sounds a bit like an advertisement for Charles Darwin Universities Bachelor of Teaching and Learning Preservice.
To do better, just say “I’m studying education, and they’re spending a lot of time on this stuff called ‘deep learning,’” and then maybe you could spend a few sentences actually talking about what you’re studying and explaining it to Robin, rather than expecting him to sign up for classes to find out.
Thanks for your response Manfred.
So it’s a “9 point” positive to say something that reiterates a commonly perceived problem and offers no solution and also makes a factual error, but to direct someone to a place that actually addresses their problem is a −1. Cool, I’m getting a feel for the website now, cheers.
Let’s try this. It’s actually pretty easy to test for deeper learning. For example multiple choice questions have previously been considered as examples of shallow learning, or if you prefer shallow testing, and in the past that was accurate and indeed in some still existing multiple choice questions there isn’t a path towards deep learning. However consider this question.
The 11 letters in the word PROBABILITY are written on 11 pieces of paper, and a piece of paper chosen at random from a bag. Which of the following statements are true? a)The probability of selecting a “B” is less than the probability of selecting an “I”. b)There is a greater chance of obtaining a consonant than of obtaining a vowel. c)A vowel is less likely than a consonant. d)If you repeated the experiment a very large number of times, approximately 63% of the results would be consonants. Make your selection, note that you may select more than zero answers.
Now since “advertising” for CDU might be deemed as somewhat negative, am I permitted to share the details of a book one could read to understand how deep learning for mathematics can be taught to middle school students? Or would that be advertising also?
Elementary & Middle School Mathematics : Teaching Developmentally by John A. Van De Walle, Karen S. Karp and Jennifer M. Bay-Williams, published by Pearson International.
If you’re Robin Hanson, sure. Also note that the comment is from ’07: the voting system wasn’t put in place until after the move from Overcoming Bias (Hanson’s blog) to LessWrong. If a similar comment was made today it would probably be voted up much higher. Or maybe downvoted, who knows.
Here, I gave you an upvote. Now it’s a 0. Karma means too little to stress out over like this.
Good day to you too.
Telling someone to read a book is still probably expecting too much of them. An online article is pretty much the limit, and it’s even better if you can tell the other person what you want them to learn in your own words. Referencing a book is good, though.
Cheers Manfred, Hopefully I’ve given a glimpse of how one can apply and test for deep learning… and if I a “lowly” first year undergrad in education can do it, seems it’s probably not that hard. [smirk]
I’m getting a few mixed signals from the site, and that’s to be expected. Some want high detail, others want little. Some will downgrade me for restating something that’s been said, others don’t seem to be downgraded for saying something that’s common knowledge. I’ll get the hang of it. I’m well aware that joining an internet site/forum/community involves some amount of fitting in, some amount of taking hits cause you’re the new guy or just aren’t getting it, and some amount of adjusting to the norms, indeed some good old fashion relationship building too.
I was hoping this might be a place where solutions are discussed more than problems are complained about. Perhaps that’s true and I just haven’t discovered it yet, don’t know.
I remain open minded and am working my way through the sequences, they’re good reading.
Disconfirmation bias—love that concept and discovered it on lesswrong, hadn’t found that one in the Social Psych or Education Psych textbook yet. Even if I get too peeved off and leave to never return, LW will be remembered fondly for that one concept alone.
Could you give an example of “shallow learning” alongside one of “deep learning,” and explain the difference? “Deep learning” definitely sounds like something that’s better than “shallow learning,” but you haven’t made if very clear what it actually is.
Desrtopa, sure thing mate. Deep learning example : The 11 letters in the word PROBABILITY are written on 11 pieces of paper, and a piece of paper chosen at random from a bag. Which of the following statements are true? a)The probability of selecting a “B” is less than the probability of selecting an “I”. b)There is a greater chance of obtaining a consonant than of obtaining a vowel. c)A vowel is less likely than a consonant. d)If you repeated the experiment a very large number of times, approximately 63% of the results would be consonants. Make your selection, note that you may select more than zero answers.
Shallow learning example. The 11 letters of FOUNDATIONS are written on 11 pieces of paper, and a piece of paper chosen at random from a bag. What is the probability that an “O” is selected? a) 3⁄12. b) 1⁄11. c) 2⁄11. d) 2⁄12. Select only 1 answer.
The Deep learning example uses the word “probability” that’s a way to prime the student to thinking in terms of probability, it is an interconnection, it enhances learning, the word “foundations” doesn’t do this. The question is “which of the following statements are true?” – that’s a question that is more open than “What is the probability that an “O” is selected?”—open questions evoke deep learning better than closed questions. The answer selections in the deep learning are worded, they require interpretation, they need an understanding of consonants and vowels – which again provides an interconnection with English, and interconnections are deep learning. Whilst the shallow learning has 4 numbers for options, they require no interpretation and they don’t interconnect with English as much as does the deep learning example. The Deep learning questions “make you selection, note that you may select more than zero answers” gives the reader a pause… how many can I select, what does more than zero mean, it requires some interpretation, could be 1, 2, 3, or 4! The “Select only 1 answer” doesn’t need interpretation, it’s closed – just 1. Now about the answers themselves. Shallow learning multiple choice questions typically have 2 options that are readily visible as incorrect and can be quickly discarded, 3⁄12 can be quickly discarded because there aren’t 3 O’s nor are there 12 letters. 1⁄11 can be quickly discarded because there are 2 “O”s. 2⁄11 is the correct answer and so the person doesn’t even need to assess answer d. Where as in the deep learning each answer needs to be assessed for their truth value, and no answer can be quickly discarded.