Novice Art Blogger, machines interpreting art


Today there’s a growing “machine audience” nurtured by projects that interpret data and use it to feed other systems (such as Google Books feeding Google Corpus for their online services). “Novice Art Blogger” by Matthew Plummer Fernandez joins this scene, reinvigorating a tradition for creating “machine-interpreted” art. This work reads images and meta-data from Tate Britain’s online archive of Abstract Art in an attempt to identify its content, automatically producing posts on a blog and using this analysis to find another ‘similar’ picture. Technically constructed around a “deep learning” algorithm and a Tumblr account, it formulates plausible visual analyses. Even if the language used is relatively simple, the recurring abstract elements and the openness of the interpretation play a decisive role in the reader’s imaginary. And upping the ante, it’d be even more fascinating to feed the posts back to textual analysis software, searching for new, accidental patterns.