[-empyre-] Ben Bogart

[B. Bogart]

Hello all,

I continue to be following the list, but not with as much consistency as 
I would like. I wanted to post a little update as I'm currently at the 
Banff Centre in Alberta, Canada (www.banffcentre.ca/) one week into a 
five week self-directed residency.

I am developing software and presentation strategies for Watching and 
Dreaming (Blade Runner) in order to contextualize the work in visual 
art. Watching and Dreaming is a series of generative video installations 
that inherit the conceptual basis of my PhD work 
(http://www.ekran.org/ben/wp/2014/a-machine-that-dreams-an-artistic-enquiry-leading-to-an-integrative-theory-and-computational-artwork/) 
where I attempted to make a Dreaming Machine.

The argument presented in the PhD work is that dreaming, mind wandering 
and external perception are contiguous and depend on shared mechanisms 
of internal simulation. What we experience of the world is the result of 
massive amounts of implicit learning where much of what we 'perceive' is 
actually the result of unconscious imagination rather than the result of 
direct access to sensation. External perception resembles reality 
because of the constraints on those mechanisms of simulation imposed by 
sensory information. In the absence of, or shift of attention away from, 
sensory information (for example during sleep or mind wandering) those 
processes continue on. No longer constrained by embodied physical 
reality they are free to diverge and converge away and toward 
plausibility. For more information see my TEDx talk 
(https://youtu.be/xYtt8qSwJws).

Whereas the Dreaming Machines learn structure from a live camera in the 
installation context, Watching and Dreaming learns from the contrived 
worlds presented in popular cinematic depictions of AI, such as Stanley 
Kubrick's 2001: A Space Odyssey and Ridley Scott's Blade Runner. The 
system uses machine learning methods (specifically clustering and 
prediction) to break film frames into millions of components and 
producing abstractions of visual information. The system 'imagines' 
perceptions by reconstructing frames using these abstractions or dreams 
by predicting which abstractions should be present in this 'train of 
thought'.

I'm blogging my learning and development process on my website 
(http://www.ekran.org/ben/wp/category/production/wd/), and you can use 
the "RSS Feed" link under "About" on the left panel if you would like to 
to subscribe to updates. I've also posted some early work in progress 
showing perceptual reconstructions on YouTube 
(https://youtu.be/AIUWu08qM2o). I expect to post more quite soon!

Beyond this I'm waiting for word on council funding to continue 
supporting this project, and I should also hear soon about my research 
council application for a postdoc in a cognitive neuroscience lab.

Thanks all for reading!
Ben Bogart