Coursera (or similar) reviews and recommendations

[chewybrian]

In case you never heard of it, Coursera is a website where you can access college courses. You can essentially audit most of them for free, or pay a small fee to get a certificate of completion, or pay more to work towards a degree.

https://www.coursera.org/

I wonder if any of you are using or have used this site or one of the others along the same lines. How did you like the experience? What didn't you like? What courses did you take? Which were your favorites and what would you recommend? How did you decide what to take?

--------------------

I am going to begin with a course called: "Mindware: Critical Thinking for the Information Age", from the University of Michigan (even though I am a Buckeyes fan!)

https://www.coursera.org/learn/mindware

My understanding is that it presents and explains cognitive biases and offers ways to allow your reason to overcome them. It seems there is no way to undo the heuristics hard-wired in the brain which are there to help us make life and death decisions quickly in the wild without stopping to think. These heuristics drive us toward faulty solutions to complex problems in the modern world which require reason over instinct. If we simply stack logic on top of the instinctual or intuitive answer, we can go horribly wrong (for example, assuming Iraq probably had weapons of mass destruction, applying the confirmation bias and favoring weak information that matched our preconception and ignoring strong information that did not match our intuition, and starting a war costing many lives and lots of capital and resources with questionable gains in the end).

If you randomly test people for biases, they tend to have them across the board. However, if you test them after explaining the possible bias, they tend to be on alert that they are being tested, and think their way past the bias and closer to the truth. If you consider that much of life turns out to be such a test in practice, then being aware that we are being tested at every turn leads us to be able to think more clearly and possibly over-ride the biases that we have.

[chewybrian]

I was able to finish this course very quickly because I had already studied logic and statistics. I was disappointed to find these were the entire focus, and there was not much if any psychology involved. For those who had not studied both statistics and logic, it might prove quite useful. I liked the professor and the content, but the content was a bit too familiar.

[Arjen]

May I inquire what kind of logic you studied?

[chewybrian]

May I inquire what kind of logic you studied?

I only took a basic class in logic 30+ years ago in college, but it left in my brain most of the principles brought forth in this class I just took, which surprised me. Much of the class time was spent on basic syllogisms, propositional logic, and simple statistical concepts like regression to the mean or the law of large numbers.

I was surprised at the professor's total disdain for multiple regression analysis. In my mind, studies of correlation were weak, but "backing out" the other factors in play should have revealed the true impact of the factor being studied. To his mind, there are way too many other factors possibly creating noise, and there is little point in trying to back them all out. Rather, there was no other proper way but to conduct a random experiment, assigning the factor to be studied by chance and allowing all other factors to even out through a normal distribution.

I was also fascinated by the contrast presented between logic and dialectical reasoning (East and West thinking if you wish), which I intend to explore further.

[Arjen]

Yeah, the Hegelian dialectic.... Much loved by dictators.

Questions:
1) What do you mean by regression to the mean and the law of large numbers?
2) As relating to what education was this logic? I am getting the impression that logic for philosophers is different. I know it differs for the purposes. So, I am curious.
3) COuld you see if you understand my distiction between sets and elements in relation to machine learning in this topic?

I think that many do this and that not doing this is rewarded. They also call this machine learning. Likelihood. Pay attention. You will see it more and more.

The problem is that every coherent thought has a major and a minor premise (and relates to another thought in some way). However, many observations used as a set doesn't form a major premise. Likelihood is not an exact determining factor.

I think this is what you are pointing to with the dialectic?

[chewybrian]

1) What do you mean by regression to the mean and the law of large numbers?

These are simple ways to help people understand how they could be misinterpreting data, and how they could understand it better by having a more reasonable expectation or interpretation. If you see a result that seems extreme, you should not form an expectation that the following result should also be extreme. Flip a coin and get 5 tails in a row, and the odds of heads or tails on the next toss is still 50/50, despite peoples' tendency to form unfounded expectations based on the small sample they just saw.

Say I have four home runs in my last 10 times at bat--great! But, if I have hit 14 total home runs in my last 1000 at bats, then rather than looking for 4 home runs in my next 10 at bats, I should be expecting zero, or one if I am lucky. That is regression to the mean.

The law of large numbers simply says that more samples or larger samples from the same set are likely to indicate something closer to the truth. The 14 home runs from 1000 at bats is a much more realistic appraisal of my power-hitting ability than the 4 home runs from the last 10 at bats, or the zero from the 10 tries before that.

2) As relating to what education was this logic? I am getting the impression that logic for philosophers is different. I know it differs for the purposes. So, I am curious.

Everything is different for philosophers. I understand logic to mean propositional logic, syllogisms, inductive and deductive reasoning, etc. Philosophers might see some other things as logic, but I don't. That doesn't mean there is no value in something because logic can not or should not be applied to the thing being discussed (like God, perhaps). It just means, to me, that I don't see logic outside of what I hold to be logic.

3) COuld you see if you understand my distiction between sets and elements in relation to machine learning in this topic?

I think that many do this and that not doing this is rewarded. They also call this machine learning. Likelihood. Pay attention. You will see it more and more.

The problem is that every coherent thought has a major and a minor premise (and relates to another thought in some way). However, many observations used as a set doesn't form a major premise. Likelihood is not an exact determining factor.

It sounds like you are describing the difference between inductive and deductive reasoning, in a broad sense. Machine learning is a method of inductive reasoning. If I send a probe from my planet to discover the nature of people on earth, it might examine various people and see their size, shape and characteristics, and be able to form some reasonable idea of what people in general are like. However, if it lands in Ireland and does all its sampling there before taking off, it may grossly underestimate the number of Chinese people, and think there are many more people on earth with red hair than there really are. So, inductive reasoning can sometimes miss the mark, but sometimes it's all you've got to work with.

There other form you are describing looks like a syllogism: main premise, observation, and conclusion.

A-All men like coffee
B-I am a man
C-I like coffee

That is a valid argument, but the premise is false (I do like coffee, though).

So, that is the weakness of deductive reasoning. We never really have a hard fact to work with in reality, but only in a theoretical world. Math is invincible within the universe of math, but it only helps us in the real world if our observations, guesses, and ideas are right on the mark. So, if I have 6 apples and give you 2, I definitely have 4 left, but... only if I was able too correctly count to 6. What if there was a smaller apple hiding behind a bigger one, such that I could not see it? Maybe I had 7 to begin with, and now I have 5.

As I see it, you can't do much thinking without both types in play. So you are correct to warn against people proceeding as if they are a computer algorithm. But, they should also not fall in love with their logic and forget how little they are able to know before stacking logic on their flimsy bit of alleged knowledge.

[Arjen]

These are simple ways to help people understand how they could be misinterpreting data, and how they could understand it better by having a more reasonable expectation or interpretation. If you see a result that seems extreme, you should not form an expectation that the following result should also be extreme. Flip a coin and get 5 tails in a row, and the odds of heads or tails on the next toss is still 50/50, despite peoples' tendency to form unfounded expectations based on the small sample they just saw.

Say I have four home runs in my last 10 times at bat--great! But, if I have hit 14 total home runs in my last 1000 at bats, then rather than looking for 4 home runs in my next 10 at bats, I should be expecting zero, or one if I am lucky. That is regression to the mean.

The law of large numbers simply says that more samples or larger samples from the same set are likely to indicate something closer to the truth. The 14 home runs from 1000 at bats is a much more realistic appraisal of my power-hitting ability than the 4 home runs from the last 10 at bats, or the zero from the 10 tries before that.

I see, I thought it was part of formal logic, but I think you just mean that those are logical thought patters. Or: How come that is included in your logical course?

Everything is different for philosophers. I understand logic to mean propositional logic, syllogisms, inductive and deductive reasoning, etc. Philosophers might see some other things as logic, but I don't. That doesn't mean there is no value in something because logic can not or should not be applied to the thing being discussed (like God, perhaps). It just means, to me, that I don't see logic outside of what I hold to be logic.

Because logic describes the working in the mind, it can be applied to all mental activities. In fact: also to observations, since those are pieced together in the mind. Therefore, logic can be applied to everything. But, Depending on the field, it is applied in different ways.

It sounds like you are describing the difference between inductive and deductive reasoning, in a broad sense. Machine learning is a method of inductive reasoning. If I send a probe from my planet to discover the nature of people on earth, it might examine various people and see their size, shape and characteristics, and be able to form some reasonable idea of what people in general are like. However, if it lands in Ireland and does all its sampling there before taking off, it may grossly underestimate the number of Chinese people, and think there are many more people on earth with red hair than there really are. So, inductive reasoning can sometimes miss the mark, but sometimes it's all you've got to work with.

There other form you are describing looks like a syllogism: main premise, observation, and conclusion.

A-All men like coffee
B-I am a man
C-I like coffee

That is a valid argument, but the premise is false (I do like coffee, though).

So, that is the weakness of deductive reasoning. We never really have a hard fact to work with in reality, but only in a theoretical world. Math is invincible within the universe of math, but it only helps us in the real world if our observations, guesses, and ideas are right on the mark. So, if I have 6 apples and give you 2, I definitely have 4 left, but... only if I was able too correctly count to 6. What if there was a smaller apple hiding behind a bigger one, such that I could not see it? Maybe I had 7 to begin with, and now I have 5.

As I see it, you can't do much thinking without both types in play. So you are correct to warn against people proceeding as if they are a computer algorithm. But, they should also not fall in love with their logic and forget how little they are able to know before stacking logic on their flimsy bit of alleged knowledge.

Yeah, it is what the Hegelian dialectic exploits.
I am glad that I was comprehensible.

[Sculptor1]

In case you never heard of it, Coursera is a website where you can access college courses. You can essentially audit most of them for free, or pay a small fee to get a certificate of completion, or pay more to work towards a degree.

https://www.coursera.org/

I wonder if any of you are using or have used this site or one of the others along the same lines. How did you like the experience? What didn't you like? What courses did you take? Which were your favorites and what would you recommend? How did you decide what to take?

--------------------

I am going to begin with a course called: "Mindware: Critical Thinking for the Information Age", from the University of Michigan (even though I am a Buckeyes fan!)

https://www.coursera.org/learn/mindware

My understanding is that it presents and explains cognitive biases and offers ways to allow your reason to overcome them. It seems there is no way to undo the heuristics hard-wired in the brain which are there to help us make life and death decisions quickly in the wild without stopping to think. These heuristics drive us toward faulty solutions to complex problems in the modern world which require reason over instinct. If we simply stack logic on top of the instinctual or intuitive answer, we can go horribly wrong (for example, assuming Iraq probably had weapons of mass destruction, applying the confirmation bias and favoring weak information that matched our preconception and ignoring strong information that did not match our intuition, and starting a war costing many lives and lots of capital and resources with questionable gains in the end).

If you randomly test people for biases, they tend to have them across the board. However, if you test them after explaining the possible bias, they tend to be on alert that they are being tested, and think their way past the bias and closer to the truth. If you consider that much of life turns out to be such a test in practice, then being aware that we are being tested at every turn leads us to be able to think more clearly and possibly over-ride the biases that we have.

There is something in the Uk called "future learn" which sounds similar.
https://www.futurelearn.com
You can follow most courses for free, if your aim is learning. They also issue certificates if you take the fee paying route.
University acreditation is hard in the UK, so I do not think they offer that. But many places such things are a bit like confetti.
One always remembers Breaking Bad's lawyer Saul with his degree from the University of American Samoa.

I have to say that in general getting qualifications remotely is liable to fraud and safegaurds are hard to establish so that employers can trust such degrees.
Nonetheless the Open Univeristy in the UK has been running for 50 years and its degrees are a match for any UK university.

http://www.open.ac.uk/courses

[chewybrian]

I'm off to Scotland to "attend" the University of Edinburgh, with a course called "Know Thyself"-The Examined Life-The Value and Limits of Self-Knowledge".

https://www.coursera.org/learn/know-thy ... mined-life

I don't have much of a preconception of what might be in this course, but my preconceptions were not that accurate last time, so I will just see what I see.

[Arjen]

Cool!
Say hello to the Oracle for me. And remember: "nothing in excess!

When you come back, tell us all ab out it

[chewybrian]

Well, the "Know Thyself" course was not that exciting. I had already read much of the material that the course used, like Plato's Apology and some Sartre. The professor was just OK, though I really don't go in for the meditation that seems to do something for him. The course was much longer than the prior course, especially if you tried to open and read al the referenced material, some of which was just comic strips and whatnot.

Meh, not much there for me.

I am staying in Edinburgh for a course called "Introduction to the Philosophy of Cognitive Sciences".

https://www.coursera.org/learn/philosop ... e-sciences

[chewybrian]

The "Philosophy of Cognitive Sciences" was the most interesting course so far, though I would say it was much about science and little about philosophy. They discussed the development of cognition and the way it evolved in response to the environment. There was a lot near the end about embodied cognition, or the way our understanding developed in ways that work with our bodies and our environments, rather than in our abstract universe of thought. They explained this relationship from the simplest organisms all the way up to people.

I was very interested to see them present an idea I've always felt was true, that our consciousness works in layers. At the highest level, we are largely unaware of what is going on at the lower levels. For example, we may be walking, but devoting little if any of our 'access consciousness' to the task. In fact, if we purposely do so, we may find we get much worse in our performance of the task. There seems to be a purposeful hierarchy at work. The lower levels carry on with the basic stuff, allowing us to put the higher levels to work on more serious concerns, like where we are going, why, and how we plan to get there.

In the middle, there was a lot of discussion about artificial intelligence, machine learning and algorithms. It is difficult to see much difference between this intelligence and our own, in the manner it was presented. The looked at the 'hard problem' of consciousness. Not surprisingly, they had no real insight to give, other than seeing that we scarcely know what questions to ask, much less what the answers might be.

I'm heading to Yale now, for a course called "The Science of Well-being". It seems to be about the science of happiness, and understanding what objective methods we might apply to ourselves to be happier in our subjective universes.

https://www.coursera.org/learn/the-scie ... me/welcome