[-empyre-] View Recent Changes
danielp73 at gmail.com
Fri Nov 20 14:45:04 AEDT 2020
thanks so much for these thoughtful answers Matt.
In case anyone wants to see some example image relations from I'm Feeling
Lucky, here are some direct links:
"appears to be" : http://im-feeling-lucky.surge.sh/#/i/appears%20to%20be
"extends over" : http://im-feeling-lucky.surge.sh/#/i/extendes%20over
"covered by a" : http://im-feeling-lucky.surge.sh/#/i/covered%20by%20a
"has many" : http://im-feeling-lucky.surge.sh/#/i/has%20many
Link to "I'm feeling lucky" on View Recent Changes:
And of course the at the accumulations exhibition:
I really love these very abstract fragments of phrases, made visually
concrete by the algorithm, which picks one set of images among infinite
I recently discussed Vannevar Bush and his Memex machine with my students.
As I'm sure many people know, the memex was an analogue device designed to
help a user make connections within and between images, texts, and other
visual records. It then recorded the process of making these links so that
each set of associations could be recalled later. On many platforms that we
use today, algorithms make these links for us based on behavioral data,
surveillance and big data. This project explores the poetic potential of
these automatic associations.
This also brings to mind https://www.are.na, the platform in which users,
rather than the platform, create their own cross referenced collections of
webpages, images and other documents. Maybe a contemporary,
visually-oriented update to some of the original ideals of hypertext.
Thanks again Matt, and everyone at View Recent Changes!
On Tue, Nov 17, 2020 at 12:54 PM Matt Nish-Lapidus <matt at emenel.ca> wrote:
> ----------empyre- soft-skinned space----------------------
> Hi empyre--
> On Nov 16, 2020, at 4:17 PM, Daniel Lichtman <danielp73 at gmail.com> wrote:
> how you see the relation between human and AI generated poetry? What
> happens when visual references that draw on our own, personal associations
> with (for example braided hair and snake curled around a stick in the entry
> for “curled”), are collected and remixed by the algorithm?
> Thanks for the thoughtful question, Dan.
> This piece started with the list of relational terms you see on the wiki
> page. I was really drawn to the poetics of it, and the complication of its
> purpose. The list of terms is used by computer vision systems to describe
> relationships between objects in images, the the terms themselves were
> written by people and compiles into this dataset.
> In the piece there’s this strange feedback loop--people define the things
> that the system can recognize as relationships, and the image search
> results tries to piece together a number of images that expresses that
> relation. The image search doesn’t actually use this type of image
> recognition system, so it looks for the relationship in metadata and we
> infer some other meaning in the collage of returned images. … which in turn
> is fluid and responding to the structure of the web via the search ranking
> algorithm. … so there are a number of layers of feedback between human
> input/activity and machine algorithms/ways of seeing. Ultimately the
> machine doesn’t know what a braid is, or even what an image is… it
> identifies patterns of pixels and metadata and supplies us with material
> from which we can try to make meaning.
> You’re right to see the relationship between this and the types of image
> assemblages we encounter daily—instagram, twitter, image search, tiktok …
> they are all different types of algorithmically aligned images (based on
> our mostly passive input as behaviour and relational metadata (what Yuk Hui
> calls a “digital object)). With “I’m Feeling Lucky” we were really trying
> to expose and mine the poetics of metadata itself, as well as the resulting
> images that are often both strange and compelling.
> empyre forum
> empyre at lists.artdesign.unsw.edu.au
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the empyre