Sabbatical Report: Project End

The goal of my past sabbatical was to conduct a project that studied how simple artificial neural networks learned about uncertain environments. This project was completed this past August 10 when I submitted a 240 page monograph for review at Comparative Cognition and Behavior Reviews. The sabbatical not only involved the actual writing this monograph, but also the collection of new results to be reported. This involved a tremendous amount of new research activity. I developed a number of new mathematical proofs about the relationship between simple artificial neural networks and Bayesian probability. I conducted hundreds of network simulations in order to collect data on how such networks behave in uncertain environments when interactions between cues serve as signals of reward probabilities. I also collected data from 200 Introductory Psychology students to measure their behavior in similar environments, and to compare human probability learning to that of my networks. There are striking similarities between network and human performance, and one of the main goals of my monograph is to use this relationship to support the claim that human probability learning can potentially be modeled by very simple artificial neural networks.

While this study of probabilistic artificial neural networks was the primary activity of my sabbatical, I have also been able to develop a new research project (in collaboration with Cor Baerveldt and his students in the Department of Psychology) on the history of the Center for Advanced Study of Theoretical Psychology at the University of Alberta. In particular, during we have explored archival materials and used our findings to explore the development of the Center’s flagship course ‘Seminar in Theoretical Psychology’ as well as the relationship of Center activities to Cold War social science. Both of these projects have led to manuscripts, one that is currently under review at History of the Human Sciences, the other soon to appear in History of Psychology. I also presented a poster at the 50^th annual meeting of Cheiron on the contents of the books in the collection of the Center which are now included in the D.E. Smith Reading Room at the University of Alberta.

Sabbatical Report: End of Month 1

Today marks the end of the first month of my current year-long sabbatical. I thought that this was as good a time as any to reflect on what I have accomplished so far, and to consider where my research is heading.

While the sabbatical is only officially one month old, the stage for the project was set in the fall term of 2016. In order to be fortunate enough to be awarded a sabbatical, one must apply for it, and part of this application involves proposing the kind of work that will be accomplished during the sabbatical. I have a long history of writing a book during each year-long sabbatical that I have been awarded; the plan for the current sabbatical was no different. I proposed using the time to draft a manuscript that extended some recent work in my lab on simple artificial neural networks and probability theory, and was lucky enough to be given the green light for this kind of project from the Faculty of Arts.

With a sabbatical plan required in the fall, it is not surprising that I was in a position to start groundwork for the current sabbatical at the end of the fall term. Much of that work has involved doing a lot of reading – since marking the final exam for my fall cognitive science course, I have read 23 books on systems theory, cybernetics, information theory, and probability. Those interested can see what I have been reading by looking through my Instagram account (https://www.instagram.com/drmrwdawson/) for pictures of covers. I use #reading to tag these posts. I have also conducted a pretty extensive simulation study (which has involved training and analyzing the performance of 500 different perceptrons) that explores how networks match the probability of outcomes in a three-cue probability learning task. In the fall, I plan to collect data from human subjects that are trained on the same task that I have used to train the networks; I am pretty excited about the main result that I expect to observe when networks and humans are compared. This has meant that I have also written programs to collect this data from humans. Importantly, I have also successfully navigated the process for getting ethics approval for this work; I haven’t collected human data for years. Most importantly, I have already crafted three complete chapters of a new book manuscript (when published, it will be my eighth book) that relate networks to probability theory and information theory, that explore the relationship between simple networks and Bayes’ theorem in probability, and that report the results of my simulations.

As August begins, the sabbatical project turns to writing the opening chapter of the new book. I have enough of a ‘feel’ for the project now that I need to put it in the context of other theories, and need to lay out its purpose, methodology, and implications. Writing this chapter, though, requires me to do a lot more reading than I have been doing. Up to this point, I have been reading a book every 10 days or so, and I have to accelerate this. In short, currently my next steps are to read, to think, and eventually to write. Some sense of the different topics that I will be considering will be appearing in the near future as Instagrammed book covers.

On Books and BS

This blog isn’t about the usual BS that my students might associate with my books – instead it is about the kind that affects my book production: blood sugar.

I have been coping with Type II diabetes since the turn of the millennium, and my control of my blood sugar levels has been sketchy at times.  When I was first diagnosed I became well-versed in typical diabetes-related problems (kidney trouble, heart trouble, eye trouble, infections and amputations).  However, during one period in which my blood sugar was out of control I discovered another issue while reading the literature at my specialist’s office.  Apparently there are a host of cognitive deficits that can occur with high blood sugar levels too.

More recently I encountered this problem first hand.  During the fall of 2013 and the winter of 2014 I was having a great deal of trouble concentrating.  I was working with an undergraduate student on an artificial neural networks and music project that required interpreting the internal structure of trained networks.  I was having a lot of difficulty making any sense of any of these networks.

Not coincidentally – although I did not realize this until later – my blood sugar had entered a phase that required stronger control.  At that time I used oral medications and an evening injection of slow acting insulin.  A visit to my specialist resulted in a new regime of pre-meal injections of fast acting insulin.  I started this treatment in the first week of April 2014.

What astonished me is that within a week it seemed as if my brain suddenly turned on.  I found that my ability to concentrate was stronger and my thinking was clearer.  On April 9, 2014 I took a look at the connection weights of a simple network that was resisting analysis, and immediately saw how the network worked.  I couldn’t believe it.  I started writing the interpretation up on April 11, in what became the first chapter of a new book.  For about a year up to that point I had a lot of difficulty writing.  After starting the new insulin regime I was working and writing daily, and a week ago submitted a new book manuscript comprised of over 300 pages, 150 figures, 50 tables, and a whole bunch of new simulation results and network interpretations.  I don’t think that this would have happened without the change in my treatment.

Of course this evidence is totally anecdotal, but I now have a lot more respect for how my blood sugar control can affect what I’m paid to do (i.e. think).  I’ll apologize in advance to my students, who will likely find more of that other BS in my new book when it comes out! I have one kind of BS under control, but have never figured out how to control the other

 

Overture: Alien Music

As described in this previous post, the  text below is a draft of one of several "interludes" to be included in a book that I am working on concerned with music and artificial neural networks.

 

Figure O-1. The musical signal composed by John Williams for Close Encounters of the Third Kind.

 

Steven Spielberg’s 1977 movie Close Encounters of the Third Kind declares that we are not alone, and we should not be afraid.  The film follows ordinary people after they experience a close encounter with an unidentified flying object.  After this experience, the protagonist of the movie, Roy Neary (played by Richard Dreyfuss), becomes obsessed with seeing the UFO again, as well as with creating sculptures (from mashed potatoes and other materials) of a shape that has a deep meaning that he cannot quite fathom.

Eventually it is revealed that this shape represents a monolith of rock (Devil’s Tower, Wyoming) marking the selected location for first contact between humans and aliens.  The obsessions of Neary and others are produced by an implanted invitation from the aliens to journey with them.  The invitees recognize the shape from television news broadcasts concerning a story about a (hoax) anthrax outbreak in Wyoming.  They spend the latter part of movie traveling to Devil’s Tower and dodging the efforts of the authorities to stop them.  In the final scene of the movie only Neary succeeds; he is welcomed by alien children as he ventures aboard their mother ship.

A simple musical signal plays a key role in the film, and is presented in the musical score above (Figure O-1).  It was composed by John Williams.  He strove to write a signal that it was long enough to be set apart from the simplest musical elements (e.g. a chord or an interval), but not so long to exist on its own as a melody.  He decided that these goals would be achieved by a theme that was only five notes in length, and composed about 350 different five-note permutaitons.  Spielberg liked the one presented in Figure O-1 the best, and it became one of the most famous musical themes in film history. 

The musical theme is first encountered in the film being sung by hundreds of pilgrims at Dharmsala, India.  When asked where they first heard it, they all point, in unison, to the sky.  In a later scene, scientists broadcast the tones into space, and receive the coordinates of Devil’s Tower in return.  At the climax of the film, the signal is performed on an ARP 2500 synthesizer located on a massive runway constructed atop Devil’s Tower.  The music greets, and communicates with, the alien visitors.

In this final scene the keyboard is played by a young, fresh-faced Phillip Dodds.  Dodds was in fact the engineer that ARP sent to set up the synthesizer used in the movie.  When the original actor who was to play it became sick, Spielberg saw Dodds working on the machine, liked his look, and cast him in the role.

As the scene unfolds Dodds is instructed to play the musical signal faster and louder while the chief scientist (Lacombe, played by Francois Truffault) strides out along the runway.  Eventually the enormous mother ship arrives, hovers over the runway, and begins to loudly copy the notes that Dodds plays.  This musical mimicing quickly erupts into an interstellar jam session of increasing tempo and complexity.  Awestruck and wide eyed, Dodds exclaims “What are we saying to each other?”

This is a very deep question indeed.

The intended message in the film is that music is – literally – a universal language, one shared by all intelligent life forms.  That the alien ship generates the same notes, and that it jams with Dodds’ character in the same musical system, supports this optimistic view.  To answer Dodds’ question a scientist standing beside him says “Seems they’re trying to teach us a basic tonal vocabulary.”  Another immediately adds “It’s the first day of school, fellas.”

However, other intriguing scenarios exist; there are alternative musical possibilities that Spielberg did not explore in his film.  Does the fact that both the ARP 2500 and the mother ship generate the same basic tonal vocabulary imply that the humans and the aliens share the same underlying musical theory?  Or instead is it possible that a completely different musical theory – an alien music theory that is dramatically different from our own – is still capable of generating the same patterns of musical notes?  Perhaps all these fellas know they are attending the first day of school, but are not aware of the lecture topic! 

In many respects, my aim in this book is to explore the possibility of such alien music.  However, instead of requiring a close encounter of the third kind for this exploration, I am fortunately able to adopt a much more terrestrial approach.

This method involves teaching a particular type of computer simulation to generate responses that are consistent with Western musical theory.  For instance, the computer simulation can be presented the tones that define a particular scale, and can learn to respond with the tonic note of that scale, or to identify that scale as being major or minor.  From this perspective, similar to the alien encounter at the end of Spielberg’s movie, the computer simulation and I are saying something to each other, and are learning to use a basic tonal vocabulary.  Our inputs and outputs are consistent.

However, the particular type of computer simulation that is being trained in this fashion is called an artificial neural network, and while it learns to generate the correct outputs to various musical inputs, it is not taught – or constrained by – traditional Western music theory.  Many researchers in what is known as connectionist cognitive science argue that the internal workings of artificial neural networks are quite distinct from the clear formal properties found in logic, mathematics, or music theory.  As a result it is possible that the artificial neural network can discover a completely novel approach – and alien music theory – that generates the same input/output relationships that are defined by Western music theory.

In order to determine whether this is possible, after a network is trained we have to examine its internal structure to discover how it is able to generate its musical responses.  As is explained in more detail in Chapters 1 through 3, an artificial neural network is a messy collection of different processors (analogous to neurons) that send signals to one another through a larger and messier collection of weighted connections (analogous to synapses between neurons).  The musical knowledge of a trained network is represented in its internal patterns of connectivity.  The messiness of such patterns – called distributed representations – is what makes the existence of an alien music theory plausible.  We can use a variety of techniques to make sense of a network’s internal structure, and in so doing can reveal the musical regularities that the network has learned, on its own, to exploit.  Of key interest is whether the theory that the network has learned is the same as our own.

Interestingly, and as will be detailed in the chapters that follow, artificial neural networks can discover novel musical theories that seem completely alien.  This has interesting implications for music, insofar as it reveals alternative musical formalisms.  This also has important implications for the study of musical cognition, because it reveals a variety of different kinds of representations that the human brain might use to process music.

 

A Prelude To Some Interludes

This blog has not been active since April, 2014, but for a very good reason: I am hard at work on a new book that describes some research that uses artificial neural networks to study music.  Last April, for a variety of reasons that may be discussed in a future blog, this research started to advance nicely.  One consequence is that I began to write book chapters -- the first 9 are in solid form, with only a couple more to go, and amount to well over 100,000 words of new prose.  Striking the book while the iron was hot meant putting other writing (aka work on this blog) on hold.

Part of my book writing involves creating short "interludes" between chapters; short pieces that highlight some general theme, and move the reader's focus away from the technical details of a particular chapter to the bigger picture.  I thought that these pieces might serve well as posts on this blog, so over the next few weeks (as more and more get created) I will test drive them here.  I will post the first (the "overture" entitled "Alien Music") here later this evening.

Writing and the Dictionary of Cognitive Science

Disseminating knowledge is one of the most important elements of my job.  One of the most effective means of disseminating knowledge is writing about what one has discovered, and publishing this writing as a journal article or as a book.

Writing, however, is an activity that requires a great deal of practice.  There are times when one cannot write during a research project – for instance, when a project is just starting or is ongoing, and you do not have any results to write about!  How do you keep your writing habits sharp in such lean times?

I try to keep in the writing habit by taking on projects that permit me to write when the need arises, but which can be left alone when other kinds of writing projects are put on the ‘front burner’.  One of my most important tools for maintaining my writing habit is the University Of Alberta Dictionary Of Cognitive Science.  It is a perfect writing project, because it can be left dormant for long periods of time.  When the need arises, it is easy to extend, because you simply find terms that can be defined, and which do not currently have a definition on the site.

 

This Dictionary of Cognitive Science is simply that: a web resource that provides brief descriptions of terms that are related to cognitive science in one way or another.  Most descriptions in this dictionary talk about a term in a paragraph or two, and provide a handful of references.  The Dictionary is organized alphabetically, and is also searchable.  It currently has entries for 373 terms, beginning with ‘abecedarium’ and ending with ‘z-lens’.  Since June 12, 2012 I have been tweeting links to its entries, in random order, every weekday as a ‘cognitive science word of the day’, or #cogsci #wotd.  I will have exhausted this list sometime by the end of 2013, and will have to add more definitions to continue this practice!

The Dictionary of Cognitive Science began as a class writing project.  Students in a graduate course on cognitive science wrote a handful of definitions, and I posted these definitions on a website that was being delivered by a lab SPARCstation.  One of my former graduate students, David Medler, wrote some routines for searching and posting on this website; David and I even published an account of this as a pedagogical activity (Dawson & Medler, 1999).  The current version is maintained as an Adobe Dreamweaver site on my office desk machine, and is served out on a new Linux box that is housed in my basement lab.

The Dictionary started as a class writing exercise, but it soon took on a life of its own.  It turned out that a lot of people were connecting to it.  As a result, I began to feel a responsibility to grow it, and to clean up entries in it that I felt were substandard.  I started this phase of the project in 2009.  In addition to editing existing entries, I added new entries related to material being moulded into a book on LEGO robots and embodied cognitive science (Dawson, Dupuis, & Wilson, 2010).  After another hiatus, another set of terms were added while I was writing my current book on the foundations of cognitive science (Dawson, 2013).

While the Dictionary of Cognitive Science exists as an internet resource, its control structure is a narrow plywood bookshelf that is slowly being filled with hundreds of index cards.  This physical instantiation of the Dictionary stands in my office, as shown in the image below.  The writing process proceeds as follows:  First, I scan through the Dictionary, looking for terms related to my current work that have no entries, or looking for terms that need to be revised.  For each term that I could create or revise, I write the term on an index card.  Second, I take a handful of these cards home, and spend a while writing an electronic version of a definition for each.  Third, I create a page for each new term in Dreamweaver, and update the required links in the website that permit the new definitions to be indexed.  At this time I stamp a ‘completed’ date on the index card of each definition that has been created.  I then upload the updated Dictionary to the server, and am ready to post the new entries on Twitter, one day at a time.  With every post, I stamp a ‘posted’ date on the index card, and then place the card on its alphabetized shelf in my office.

 

 

As the fall term comes to an end, I will have run out of current Dictionary entries.  Fortunately, I am working on a couple of new projects (relating neural networks to Bayes theorem, and a book about interpreting the internal structure of neural networks trained on musical tasks), which will serve as a source of (needed) new material for the Dictionary of Cognitive Science.  I am stocked up on index cards, and ready to write new definitions to keep the rust off my writing while I collect new data on these new projects.

References

 

 

·         Dawson, M. R. W. (2013). Mind, Body, World: Foundations of Cognitive Science. Edmonton, AB: Athabasca University Press.

·         Dawson, M.R.W., Dupuis, B., & Wilson, M. (2010). From Bricks to Brains: The Embodied Cognitive Science of LEGO Robots. Athabasca University Press, Edmonton.

·         Dawson, M. R. W., & Medler, D. A. (1999). The Dictionary of Cognitive Science: One approach to teaching students how to create their own WWW instructional materials. International Journal of Educational Telecommunications, 5, 65-78.

 

 

 

Eight Tips About Publishing Books

This past Friday marked the release of my new book, Mind, Body, World: Foundations Of Cognitive Science, published by Athabasca University Press.  The book, which is described (and – shameless plug – can be ordered) here, is my attempt to introduce cognitive science, and at the same time consider the relationships between three different schools of cognitive science: classical, connectionist, and embodied.  The book is part of Athabasca University Press’ OPEL(Open Paths to Enriched Learning) series.  This series is an attempt to reduce barriers to university access, in particular the cost of course materials.  The book is available for a very reasonable price in both paperback and e-book format, and a free pdf version of the book is also available from the publisher. Enough with the shameless plugs for the book!  I am looking forward to seeing it in its physical form, which I am hoping will be possible next week – because that is when some students might like to buy it to use as a text for my third-year “Foundations of Cognitive Science” course.

With the release of this latest work, I thought that I would celebrate by sharing what I have learned about the process of writing a book, and then getting it published.  Mind, Body, World is my sixth book; three have been published by Wiley-Blackwell, two by Athabasca University Press, and one (a really cool study of the relationship between connectionism and classical conditioning) has appeared as a monograph in Comparative Cognition & Behavior Review.  Hopefully I have learned something about the process with all of this practice!

Tip #1: Write a book because you feel that you need to.  Writing a book is a great deal of work, and you can spend a lot of time working on a manuscript without any immediate reward.  In short, in the ‘publish or perish’ world of academia, book writing can be a dangerous activity.  So, write a book because you have a strong need to create a more in depth manuscript, because you are both interested and committed to the project, and because you want to learn a great deal about the domain that you are going to be writing about.

Tip # 2: Write a complete draft of the book before you do anything else (like contact a prospective publisher.  With one exception, every book manuscript that I have written was in the form of a solid, complete draft before I began shopping it around to publishers.  Acquisition editors have expressed puzzlement over this policy of mind; they are willing to consider offering a book contract on the basis of a prospectus and a couple of draft chapters to be sent to reviewers.  However, I have found that complete drafts are useful for a number of reasons.

One is that publishers can be whimsical.  For instance, I knew before I started my last sabbatical that I was going to write another book – see Tip #1 – and I had a nice chat in my lab with an acquisition editor from a major publisher who was interested in it.  I told them (in violation of this tip) that I would consider sending them a prospectus and the first five chapters when I was happy with them.  Before these chapters were written, there was a change in acquisition editors, and when I was ready to send a part of the manuscript for review, they didn’t even want it!

Another is that the publication process can be fairly lengthy, because it involves one or two stages of manuscript review, creation of a document (and figures) in formats that are used by the publisher, copy editing, proofing, index creation, and the like.  While all of this is going on, you can test your completed book (Tip #2) in your course.  I did so in Fall of 2012, and used student feedback to try to make parts of the new book make more sense.  Students also found lots of copyedit issues that I was able to pass on to the publisher.

Tip #3: Write your prospectus – a brief summary of the book, its goals, its intended audience, with chapter summaries – after a full draft of the book is completed.  It is easier this way, particularly if you buy into Tip #2.  Also, I don’t write well from outlines – having a prospectus in hand does not help me much with the writing process.

Tip #4: Be patient when you shop your manuscript around.  With a complete draft of the book, as well as a prospectus, in hand you are ready to contact a publisher to determine whether they are interested in pursuing your project.  You might be told nearly immediately that they are not interested for a variety of reasons (Tip #4a: Develop a thick skin!).  If they are interested, then they will send the manuscript out for review.  This will take a while, and after this process they still might decide not to offer a contract (Tip #4a).  It can take a long while to find a publisher who is interested in bringing your manuscript to life.

Tip #5: Take reviews to heart.  If you do get a set of reviews back, take them to heart.  They are not personal attacks (usually – not true for my first book!), and by considering them objectively and constructively you can improve your manuscript.  This is particularly true if your manuscript has been rejected.  Indeed, you should not only revise your manuscript, but pass the reviews (and how you used them) on to the next publishers.  The more information a publisher has about your manuscript, the better they are able to make an informed decision about taking your project on.  Publishers are usually interested in, and thank you for, reviews that you received from other publishers that you have contacted.

Tip #6: Do not trust contracts.  Just because you have signed a contract with a publisher does not mean that they will produce your book.  True story: I signed a contract with a major publisher for my first book, and then spent six months revising it on the basis of a lengthy review that they gave me.  They even gave me a sizable advance on the manuscript.  As the revising proceeded, I sent things to them, but did not receive any replies, which puzzled me.  Eventually, they did send me a letter: telling me that the fellow who had signed me had died, that they were not going to publish the book (the contract provided an out for them), and that I needed to repay the advance.  After a year of legal hassles, I paid them back half of the advance in exchange for the rights to the book (which they maintained because of the signed contract) in order to sign a contract with a different publisher.

Tip #7: Do not accept an advance.  See Tip #6.  I have only been offered an advance once, and was foolish enough to accept it.

Tip #8: Be committed and confident.  Finding a publisher can be a challenging and frustrating and time consuming process.  I have been lucky enough to have been able to get all of my book projects ‘shipped as product’.  However, it has not always been easy.  Surviving the period of shopping a manuscript around requires that you feel strongly that the project is good and worthwhile.  This is why Tip #1 is probably the most important one in this list!

Happy writing – and publishing!