Skip to content →

Tag: psychology

Learning curves and modelling in machine learning

In this post, I am going to describe what I have just learned from Andrew Ng at Stanford about “learning curves”.  To computer scientist, a learning curve is what you might expect but describes how well data has been modeled.

I write this as a classically trained psychologist and it is clear that if we are to understand machine learning, we have to watch out for where the thinking of computer scientists differs radically from our own.  This is my commonsensical comparison of the two approaches.  I am writing it down to make sure I have followed what I heard.  It is rough and ready but may help you understand the differences between the two disciplines.

A learning curve in CS

Simply, the CStists take random samples of data where the first sample is very small, let’s say 1 because that is helpful to understanding the logic, and the last sample will be large, let’s say a few thousand.  This is random samples from the same large data set.

Generally, with a sample of 1 up to 3, we can model perfectly.  However, when we try the same model with another sample of the same size, the model will not predict well at all. The amounts of error for the experimental sample and the comparison sample will be hugely different.  So far so good. That’s what we all learned at uni.  Modelling on a small sample is the equivalent of an ‘anecodote’.  Whatever we observed may or may not transfer to other situations.

As we increase our sample size, paradoxically the amount of error in our model increases but the amount of error in our comparison situation decreases.  And ultimately, the error we are making in the two situations converges.  We also know this from uni.

Much of our training goes into getting us to do this and to increasing the sample size so that the error in the hypothetical model goes up, and the error in the comparison model goes down.  Plot this on a piece of paper with error on the y axis and sample size on the x axis.

When the two error rates converge, that is we can explain the future as well as we can explain the present, then we stop and say, “Hey, I have found a scientific law!”

I would say that our willingness to tolerate a more general description of a particular situation so that we can generalize at the same level of accuracy (and inaccuracy) to another situation is one of the hallmarks of uni training. This is so counter-intuitive that many people resist so it takes uni training to get us to do it.

What the computer scientists implicitly point out is that the converse is also true. We are now able to explain the future as badly as we explain the present!  They call this underfitting and suggest that we try another model to see if we can do a better job of explaining the present.  So we will stop increasing the sample size and start playing with the model. We can vary the form of the model, typically moving from a linear to a non-linear model (that is adding more features) and increasing the weights of the parameters (go from a loose floppy kind of model to a stiffer model, if you like).

They do this until the model overfits. That is, until our explanation of the present is very good but the same explanation produces errors in comparison situations.  When they reach this point, they backtrack to a less complicated model (fewer non-linear terms) and decrease the weights of the parameters (take note of a feature but not put too much emphasis on it.)

Once they have found this happy middle ground with a more complicated model, but without the expense of collecting more data, they will try it out on a completely new set of data.

Break with common practice in psychology

For any psychologists reading this

  • This kind of thinking provides us with a possibility of getting away from models that have been stagnant for decades.  Many of these models predict the present so-so and the future so-so.  Here is the opportunity to break away.
  • Note that machine learning specialists use procedures that look like statistics but abandon the central idea of statistics.  They aren’t promising that their original sample was randomly chosen and they aren’t directly interested in the assertion that “if and only if our original sample was random, then what we found in the sample generalizes to other samples that have also been chosen randomly”.  Though they do something similar (taking lots of randomly chosen slices of data from the data they have), they aren’t in the business of asserting the world will never change again.  They have high speed computers to crunch more data when it becomes clear that the world has changed (or that our model of the world is slightly off).
  • Many of the rules-of-thumb that we were once taught fall away. Specifically, get a large sample, keep the number of features below the size of the sample, keep the model simple – these prescriptions are not relevant once we change our starting point.  All we want to find is the model that can generalize from one situation to another with the least error and high speed computers allow us both to use more complicated models and recomputed them when the world they described changes.

I am still to see good working examples outside marketing on the one hand and robotics on the other, but it seemed worth while trying to describe the mental shift that a classically trained psychologist will go through.  Hope this helps

CHECK OUT SIMILAR POSTS

Leave a Comment

5 pack of slides on complexity theory, leadership, management and psychology

I prepared these slides to help students revise and consolidate what they know about complex adaptive systems, psychology, management, leadership, organizational design, supply networks, and yes, happiness.

View more presentations from Jo Jordan

Leave a Comment

3 simple ideas for leading in today’s turbulent workplaces

Walk-this-way-by-garryknight-via-Flickr.jpgDespite waves of change, life stood still

We were at war

I went to university at a time of radical social change.  Not to put too fine a point on it, we were in the middle of a revolutionary war.

But psychology was cruising a plateau

My nation might have been at war with itself, but profession was not undergoing great change.  Being a student was a matter of learning about behaviorism and functionalism and Marxism and   .  .  . and  .  .  .

And psychology continued to cruise even when real change had happened

It was only later that cognition made sufficient impact to affect professional life and one look at textbooks will tell you that psychologists were so complacent about the permanency of their approach that they simply edited cognition out of the applied text books.

An astonishing number of people have been left behind

Wave after wave of students have graduated without knowing how to do cognitive task analysis and if they have a glimmer, they do cognitive task analysis without agency.

If you believe the typical psychologist, people do work without knowing what they are doing or caring about what they do.

Mindfulness means the story of here & now

Students don’t even study management because organizations “just are”.    It doesn’t even seem to occur to psychologists that context is king.  Mindfulness does not seem to suggest that paying attention to the moment may be important because the moment is important.  We look for generalizations because we believe that generalizations hold and following perfect recipes is the formula for the good life.

How deadening.  How certain to create depression and ill health.  How certain to lead to economic and financial disaster.

How odd.

From paying attention to action in the moment

If

  • Context
  • Attention
  • Visualizing (not planning) and getting feedback (not making assumptions)

are both better descriptions and prescriptions of life at work, then what are the actions?

Last night, I read Gail Fairhurst’s paper on new ways of understanding leadership.  She describes new ways of thinking about work.

  1. “Delve deep into context” and be content with understanding all the different ways that the people present understand and talk about the issues.
  2. “carve out room for maneuver while others remain stymied by disparate or oppositional Discourses (Huspek, 2000)”
  3. “draw upon alternative Discourses” to have fun

OK, the have fun bit was mine.  But, the remainder of the 3 points are from Gail T. Fairhurst.

This is very different from the psychology and management of my youth which assumed:

  • There is a good way to do this
  • The old guard know best
  • This is what you have to do (and please leave mind, spirit and sense of humour at the door!)

To recap: The action of  here & now

  1. The truth is in the wide range of realities described by people who are present.
  2. Some views  will be mind-hoovering, locked in old conflicts and defining the world as impossible.  Find the way forward.  There always is one.
  3. Present (and act out) alternatives in a spirit of fun.

In thirty years’ time, people may think differently again. And so they should.  What counts are the views of people who are there at the time!

 

Resources

Fairhust, G. T. (2009).  Considering context in discursive leadership research. Human Relations, 62(11), 1607-1633.

Download a copy fast because Journals don’t give away freebies all that often.

One Comment

Anton Wilhem Amo – First African thesis in psychology

Vuvuzela Up close and personal by voltageek via FlickrCan you guess the date of the first African thesis in psychology?

I’ll tell you that it is really early – well before Wundt.

Have your written down the date?

Now you can download (from the files below) the Life and Times of Anton Wilhem Amo who wrote the first known thesis in psychology by an African.  It’s a 2 page Word document.

I guarantee you will be surprised.

4 Comments

3 ages of control

Leaving adolescence

It’s interesting when we start to take control of our lives.  We make a plan.  Sometimes it works; sometimes it doesn’t.  And we resign ourselves to being powerless.

Encountering adulthood

Then we get a bit older and we resolve to make things work.  And we do. When a plan threatens to come apart, we jump around and keep it altogether.  And feel very good for it.

Muddling through middle age

It’s only much later that we realize that we weren’t really keeping things together. We were feeling better. We were exploring other stories about ourselves in the world.

Not confronting the experiences of middle age

I see the converse too.  I know people who are brilliant at retelling a story as if the world does it’s bidding.  They can’t countenance a notion that sometimes the world really is not on your side.

They’ve never made the transition from that early stage of needing to be in control.  They’ve just learned to divert their strong need to be in control to a story that convinces .  .   . well, them.  It doesn’t convince anyone else. They are still aiming to feel better and they are willing to pervert reality to regain that feeling.

Living honestly with our lack of control

I can’t believe that this self-deception is a good thing.  Misreading the world is dangerous.  The world simply doesn’t do our bidding.

Our best bet is to position ourselves in the river and go with the current, steering lightly but not fighting.   It’s tough though. I still don’t like being washed along.  I have to reverse attitudes I worked so hard to learn.

But maybe I can achieve more through inaction?

There!  I still want to achieve.  Maybe by promising myself that prize, I can experiment with inaction and simply enjoy the river in all its tumultus chaos?

Leave a Comment

Be my dictionary. What is the difference between earnest and sincere, solemn and serious?

Speaking properly

I once worked with people who hated other people ~ or so they said.  I hated broccoli (I like it now) and I hated doing my tax return.

Sometimes I loathed someone.  Or I disliked someone.

Speaking accurately

The nuances of emotional words are interesting.  We have poor emotional vocabularies as a general rule.

Understanding nuances

The other day, I dipped into  Kate Fox’ Watching the English: The Hidden Rules of English Behaviour.

The nuances of Englishness

Kate Fox suggests the English accept serious but not solemn and sincere but not earnest. I looked up the differences in my COBUILD dictionary that was built from a corpus of actual English usage (Collins Birmingham University International Language Database).  It wasn’t very enlightening.

Serious not solemn; sincere not earnest

So I am on the trail to distinguish solemn from serious and earnest from sincere.

Any suggestions?

2 Comments

Walking with the elephants: remembering Galba Bright

Galba Bright

Many of you will remember Galba Bright. The British Sierra Leonian migrant to Jamaica who built a successful emotional intelligence website in less than a year.  He died very suddenly and many of us miss him.

Shortly before he died, Galba set me a challenging questions. Do “in tune” people reflect?

When Galba died, I had two unfinished posts on my computer.  They’ve stayed here for quite a while and a Twitter poll urged me to publish them in tribute to a man who many of us found inspiring.

This is the draft that I find the more inspiring.

Walking with Elephants

zimbabwe_hwange_national_park-compressed.jpg

Galba Bright of TuneUpYourEQ asked me to expand my comment that people who are tuned into the world don’t reflect much.  I thought this picture of Paul Van R bicycling in Zimbabwe illustrates the point I wanted to make.

Of course, we laugh at first.  Then we may wonder whether Paul was being slightly reckless.  We question his good sense and  wonder if he knows what he is doing.

If he does know what he is doing, if he understands elephants, if he knows when they are likely to walk on the road, if he knows how they will react when they see him, then he is not necessarily reckless at all.

Moreover, if he meets an elephant and the meeting is cordial, if the the elephant was allowed to be an elephant and do elephantly things in an elephantly way, then that evening Paul is likely to relax with some fond and pleasant memories.

Of course, if he doesn’t know much about elephants and he reacts to any elephants he meets in a ways that elephants don’t much like, he might spend the evening in a whole different form of reflection.

We could flesh out this question quite a lot more.  I thought it would be fun though to think about elephants.

I think my point is that when we are “in tune” with the world, we don’t reflect very much. We are connected. We are in touch.  We are enjoying the world and ‘dancing’ with its rhythms.

When we are not “in tune” with the world, then it is time to reflect. Then it is time to focus on where we are in touch, where we feel vital and alive, and what to follow and do more of.

And as most days are not blissful rides through Africa on a hot, sultry day, some time spent each evening in reflection and when we awake in the morning, helps keep us in touch with what keeps us in touch.  Some reflection calms down our fretful helter-skelter rush into stressful activity that is poor replacement for what we love.

We miss you, Galba.

Enhanced by Zemanta
Leave a Comment

Oh, yes. Exactly what I was looking for. Exploding flavors of life.

Mysteries, Yes

Truly, we live with mysteries too marvelous

to be understood.

How grass can be nourishing in the

mouths of the lambs.

How rivers and stones are forever

in allegiance with gravity

while we ourselves dream of rising.

How two hands touch and the bonds

will never be broken.

How people come, from delight or the

scars of damage,

to the comfort of a poem.

Let me keep my distance, always, from those

who think they have the answers.

Let me keep company always with those who say

“Look!” and laugh in astonishment,

and bow their heads.

Mary Oliver, Evidence

Dedicated to psychologists everywhere.

Enhanced by Zemanta
Leave a Comment

Thinking out my 2 steps for building early stage forums

Get specific help fast on the internet

Ask specific questions on Linkedin

A long time ago, I asked a very specific question on Linkedin.  “How do we find a ship at sea?”

Through the weekend, insurance professionals and ship’s captains coached me on how to deal with Lloyds, how to track the vessel’s responders, and even how to detect illegal changes in a ship’s registration number.  It was great.  Real experts stepped up and coached me on important details.

Present the pertinent facts to fellow experts on Stackoverflow

Other examples of great information exchange also take place on the internet.  Stackoverflow is a forum where computer professionals also pose specific questions usually as a short paragraph stating what happened and what they have already tried.   Members give each other advice and rate both the answers and the questions.

Stackoverflow platform and other communties

Stackoverflow is rather a famous forum because it works.  Late last year, they released their code as a white label so that other communities can use it as platform for their specialist forums.

Stackoverflow rarely works as well though for other communities though.  This seems to be the reasons why forums “fail”.

3 reasons for “forum fail”

#1  People come to “chat” rather than to ask questions.   Their goal is to punt for clients or to gain some kind of nebulous networking status.

#2  The situations are not “important”.  The only person who benefits from the answer to the question is the asker (if they even asked a genuine question).

#3  The situations are not “hard”.  The situations are ambiguous and uncertain, to be sure.  But they ae not “hard”.  “Hard” situations involved double or treble loop.  The basic question in a “hard” situation is “Is it me or should it be this hard to do?”  Forums do not do so well on triple loop learning where we are asking how, how hard AND whether it is important.

3 guides for “winning forums”

To turn these problems around, forums might succeed when

#1 There are genuinely “important” problems.  That is, there is something I must do to to help a customer.

When I am in IT and I am trying to get IT running to support an entire organization, that is important.  Figuring out how to please my professor, on the other hand  is not “important”.  It has not importance outside of itself.

#2  The situations are “hard”.   That is, it is hard to tell if we are making a mistake or if the solution is not possible.  In these conditions, a expert coach helps our learning curve enormously.

Stackoveflow asks its members to ask questions that can be answered rather than “discussed”.  Most forums are dominated by questions that are broad and vague, or, they are not specific about what the asked is trying to achieve and what they have tried so far.

#3  A community of expertise already in exists.  A community of people have very similar problems to solve and shared ways of attempting solutions.

Far too often, questions on forums require a thorough audit of context, resources and skills.  That is, they require intervention of a professional to ask what is required here, what resources do we have and what skills do we have.  On Stackoverflow, professionals are speaking to each other, and the learning task is to cope with double loop learning (and render it single loop by providing help, coaching and support).

Thought experiments about forums that work and forums that don’t work

To test this trio of criteria, I’ve thought up a two forums and used the criteria to think about when and how the forums would work

A cooking forum

# 1 Importance

I am cooking for someone else.  I want to them to enjoy their meal.

#2  Hard

I can search endlessly for the “right” recipe with locally available ingredients.  Or, I can take directions from someone who has already located available ingredients and isolated what can be done with them.

#3  Community

There are other people who cook for similar social situations with similar ingredients.  I have lived in communities who don’t put a high premium on cooking well.  Cooking has a limited range of ingredients and skill, and quantity is more important than quality.   In these communities, there is no call for a cooking forum.  Equally, in a large city with many single people, there may be call for a community interested in the best places to get good food.   It is simply not cost effective (or pleasurable) to cook pizza, English breakfast, dim sum, and so on for one person in a tiny apartment.

A forum for professional psychologists

#1 Importance

Do we have a common understanding of our customers?  Clincial and educational psychologists might have a common understanding, but do we in work & organizational psychology?  Do HR managers have a common understanding of their customers?  Hmm . .

#2 Hard

In my experience of small groups of work & organizational psychologists who trust each other, “hard” questions usually hinge on engaging the customer.   In groups who do not trust each other, technical questions are usually a proxy for the question – how do I frame the issues for these customers (who I shall not name either because I don’t trust you or because I am embarrassed by my lack of know how.)

#3 Community

Is there a community of work & organizational psychologists who are commited to this project of understanding their customers.  Or are we, like my cooking community, who only worried about eating more?    Are we competing with each other rather than collaborating on the common project of solving problems that are important and hard?

How to develop a community before we launch a forum

So what do we do when it seems that our forum sucks because we don’t have a community?

# [Hard] Answer our own questions.  Write a blog?  Writing clarifies thinking – it does for me, anyway.  That is what I am doing now.

# [Important] Clarify the social situation of our question, at least for ourselves.  For example, single well-off people in a big city don’t want to cook.  But they do want to eat well.  We can write about what they want to know, which is where reasonably priced food is exquisite. Work & organizational psychologists don’t know much about their customers.  So write about customers and concerns from the customers’ point of view being quite clear which elements of psychology were useful to them (and which were not)?

# [Community] In a previous incarnation, I was able to use institutional means to develop a professional community.  We developed programmes.  We developed alliances.  We arrange mutual continuing education by doing a ‘stretch’ project together each year.  We collaborated to create face-to-face sessions that linked noobes to experts, noobes to noobes, and experts to experts.

When we don’t have institutional resources at our disposal, we can use the internet to put ourselves out there.  Web2.0 facilties like blogs, Slideshare Linkedin, Twitter, Yahoo Upcoming and Dopplr help people find us.  So do contact forms on our websites and blog technologies that allow people to comment directly onto our website pages.  The apex of “2.0”, I think, is aranging meetups and hackdays to bring people together to develop mutual projects.  Well that was our mutual annual CPD project – those were fun days.

I haven’t seen many stories of people who have developed communities from scratch.  It is improbable anyway.  Because we must share an “important” concern and believe it is so “important” that we will help each other to navigate “hard”.  So we are looking at latent communities.

  • Step 1.  Begin with the customer who we don’t serve well, perhaps.  Not the customer who is a cash cow.  But the customer who needs something done and whose lives would be so much better if it were done.
  • Step 2. Then use Web2.0 technologies to allow people who are searching for answers to find you.

And document it!  We need more stories.

Comments?

Build the roads to bring people to you.  Host the conversation, in other words, but don’t expect the success of Stackoverflow at the outset.

Leave a Comment

Paolo Coelho on happiness and two challenges for psychologists

Psychologists need poetry

I have one piece of advice for anyone who aspires to be a psychologist.  Read poetry.  Read good novels.

Your College or Department will jump your through a  lot of pseud-scientific hoops.  Jump through them but for a different reason to the one they give.  Jump through them because they will teach you how to ‘fail informatively’.  Yes. Fail informatively.

In the future, you will be able to handle unfamiliar situations by proposing one or more reasonable ways forward.  And then you can set up some experiments.  You can choose the best way forward.  And if you have set up your experiment well, the less favorable ways will also teach you a little more than ‘wrong way’.  This is the reason why you should study science.

To understand people, well, meet a lot of people and do things with them.  And read.

A good read is Paolo Coelho who also blogs and tweets.  Today he posted a 1 minute parable on the meaning of happiness.  It is an easy read.  The ending sums up the meaning of happiness.

For psychologists out there, this parable talks about two important psychological phenonena.

#1  Management of attention.

To manage one’s own direction and to pay attention to what is going on around us.

We need lots of practice at doing this. Computer games help us do this.  TV and reading books does not.  Sport helps us learn this.  Writing does not.  But speaking does.  Make sure you get lots of practice at learning to manage your attention so that you tackle frontiers with greater ease!

#2  We live at our frontier.

To define who we are by what we do.

Not what we feel, or believe.  But what we do in various contexts defined by who else is there.  We are our frontier.  We are our edge.

Perhaps we are a young man who cannot carry two drops of oil and look around a new place.  Or frontier is the new place, the new idea, and our own confusion.  It is here that we are ‘alive’ with our dreams and our hopes, our confusions and our sorrows.

This is a tough challenge for psychologists.  We have nothing to measure.  The definition may even be circular.  That is because psychology is not a thing. It is a goal or a purpose that is supremely personal.  Our goal is to live a our frontier.  The story of our frontier and our confusion is the story we all want to hear.

When we want to do the maths, then we look at whether we were in a situation that covers the whole gamut of emotions and whether we were able to respond appropriately as events unfolded.  Or were we like the young boy, first forgetting the context and then forgetting his task.  Can we recover from confusion and distress or do we get stuck?  Are we so scared of life that we insist that it be plain sailing all day and every day?

Do we approach our frontier or do we hang back?  And under what conditions are we able to approach our frontier and learn to carry the oil and look around despite our initial confusion?

Yes, positive psychologists do know something about this.  But so do poets.  Begin with them.

Enhanced by Zemanta
2 Comments