CAN: Creative Adversarial Networks, Generating Art by Learning About Styles and Deviating from Style Norms

CAN: Creative Adversarial Networks, Generating Art by Learning About Styles and Deviating from Style Norms Ahmed Elgammal Bingchen Liu Mohamed Elhoseiny Marian Mazzone

Tag layer of the DHSS Meta-Tagging Project. The article text is reproduced from the staged copy named in sourceDesc and is not altered; the tags are held separately in standOff so that overlapping and nested spans are preserved.

CAN: Creative Adversarial Networks, Generating Art by Learning About Styles and Deviating from Style Norms 2017 ICCC / arXiv 1706.07068 elgammal2017-creative-gan elgammal2017-creative-gan.txt, the project's staged copy b663c2afcfc3874455fec84cd3b9b55db386a2db arxiv-open art-history

the paper's own words

the tag layer; a sentence written by the tagger about the paper, always anchored to a quotation but not itself a quotation

0 spans located by searching the stored text for the quote's exact words, allowing any whitespace between them method Quotation withheld because this record is not cleared for redistribution. machine generation of creative visual art Quotation withheld because this record is not cleared for redistribution. system-or-tool-building system-or-tool-building Quotation withheld because this record is not cleared for redistribution. mechanisms modify the GAN objective to push generated images to deviate from established style norms while staying near the art distribution Quotation withheld because this record is not cleared for redistribution. GANs have a fundamental limitation in generating genuinely creative works what this limits: the generative approach itself: Quotation withheld Quotation withheld because this record is not cleared for redistribution. Creativity of machine-generated artifacts is acknowledged to be an unresolved assessment problem. what this limits: every creativity claim in the paper: ASSESSING THE CREATIVITY OF ARTIFACTS GENERATED BY THE MACHINE IS AN OPEN AND HARD QUESTION Quotation withheld because this record is not cleared for redistribution. Creativity was operationalised only as novelty, because influence could not be measured. what this limits: what the word creative means in this work: SINCE INFLUENCE IS NOT RELEVANT HERE, WE USE NOVELTY AS A PROXY FOR CREATIVITY Quotation withheld because this record is not cleared for redistribution. The system has no semantic understanding of art beyond style categories. what this limits: any semantic reading of what the system does: IT DOES NOT HAVE ANY SEMANTIC UNDERSTANDING OF ART BEHIND THE CONCEPT OF STYLES Quotation withheld because this record is not cleared for redistribution. The authors cannot explain why subjects rated CAN images above the human Art Basel set. what this limits: the headline comparison, which the authors decline to interpret: WE LEAVE OPEN HOW TO INTERPRET THE HUMAN SUBJECTS' RESPONSES THAT RANKED THE CAN ART BETTER THAN THE ART BASEL SAMPLES Quotation withheld because this record is not cleared for redistribution. Further experiments are needed to settle the interpretation of the results. what this limits: the same scope as limitations[4]: MORE EXPERIMENTS ARE DEFINITELY NEEDED TO HELP ANSWER THESE QUESTIONS Quotation withheld because this record is not cleared for redistribution. Creativity is implemented as making a classifier maximally unsure which style it is looking at. what this limits: the creativity claimed for these images: it is implemented as maximising the entropy of the style class posterior, that is, as being hard for the style classifier to assign to any class Quotation withheld because this record is not cleared for redistribution. The style vocabulary is a fixed list whose length is a design parameter of the network. what this limits: the style categories used here: they are a fixed set whose size K is a network design parameter, fixed by the label set of the training corpus rather than by art history Quotation withheld because this record is not cleared for redistribution. The paper's own object is scare-quoted in its title. what this limits: the claim that this system generates art: the paper's title puts the word in quotation marks and describes the method as learning about styles and deviating from style norms Quotation withheld because this record is not cleared for redistribution. generative adversarial networks adapted to generate art that deviates from style norms Quotation withheld because this record is not cleared for redistribution. Berlyne theory of aesthetic novelty and arousal used to design the creativity objective Quotation withheld because this record is not cleared for redistribution. an art-trained DCGAN baseline at both 64x64 and 256x256, generated and rated alongside the proposed model by the same human subjects eliminates: the human ratings reflecting that the images are art-like at all rather than that the proposed objective did anything - the baseline is the same architecture without it · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. style style Quotation withheld because this record is not cleared for redistribution. style class style class Quotation withheld because this record is not cleared for redistribution. style ambiguity style ambiguity Quotation withheld because this record is not cleared for redistribution. collative variables collative variables Quotation withheld because this record is not cleared for redistribution. novelty novelty Quotation withheld because this record is not cleared for redistribution. style - AS A CLASS IN A CLASSIFIER: the network's final layer has K outputs where K is the number of style classes, so here the word names a position in a label set rather than any property of the picture, and an image's style is whichever index wins Quotation withheld because this record is not cleared for redistribution. style ambiguity - AS POSTERIOR ENTROPY: the generator is rewarded for making the style classifier uncertain, so 'deviating from style norms' is implemented as producing images the classifier cannot place, and creativity becomes a property of a confusion matrix Quotation withheld because this record is not cleared for redistribution. novelty - INHERITED FROM BERLYNE'S COLLATIVE VARIABLES: it is one of five arousal-raising properties named in an aesthetics theory - novelty, surprisingness, complexity, ambiguity and puzzlingness - and this system implements the fourth of them while the record's limitation says creativity was operationalised as the first Quotation withheld because this record is not cleared for redistribution. extends → Generative Adversarial Networks (from AI) Quotation withheld because this record is not cleared for redistribution. builds-on → Generative Adversarial Networks (GAN) Quotation withheld because this record is not cleared for redistribution. extends → Generative Adversarial Networks (GAN) Quotation withheld because this record is not cleared for redistribution. builds-on → Martindale's principle of least effort and explanation of style breaks Quotation withheld because this record is not cleared for redistribution. builds-on → the art-perception experiment of Snapper et al [23] Quotation withheld because this record is not cleared for redistribution. style ambiguity —operationalised-as→ style class style ambiguity style class strength of the relation: 1 Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but ambiguity is achieved by maximizing the style-class posterior entropy; operationalised-as is loose. Quotation withheld because this record is not cleared for redistribution. novelty —part-of→ collative variables novelty collative variables strength of the relation: 1 Quotation withheld because this record is not cleared for redistribution. Experiments compared human subjects' responses to the generated art with their responses to art created by people. Quotation withheld because this record is not cleared for redistribution. arousal potential arousal potential Quotation withheld because this record is not cleared for redistribution. Creative Adversarial Network (CAN) Creative Adversarial Network (CAN) about this quote: QUOTE REPAIRED: the extractor split a hyphen-space and the quote was recorded faithfully with the damage in it, which made it unanchorable. The repaired string is an EXACT SUBSTRING of the staged text - that is why it was accepted, not because a rule produced it. The text as originally recorded is kept in evidence as extracted. Quotation withheld because this record is not cleared for redistribution. collative variables collative variables Quotation withheld because this record is not cleared for redistribution. Wundt curve Wundt curve Quotation withheld because this record is not cleared for redistribution. hypothesis-testing hypothesis-testing about this quote: QUOTE REPAIRED: the extractor split a unicode and the quote was recorded faithfully with the damage in it, which made it unanchorable. The repaired string is an EXACT SUBSTRING of the staged text - that is why it was accepted, not because a rule produced it. The text as originally recorded is kept in evidence as extracted. Quotation withheld because this record is not cleared for redistribution. null-hypothesis null-hypothesis about this quote: QUOTE REPAIRED: the extractor split a unicode and the quote was recorded faithfully with the damage in it, which made it unanchorable. The repaired string is an EXACT SUBSTRING of the staged text - that is why it was accepted, not because a rule produced it. The text as originally recorded is kept in evidence as extracted. Quotation withheld because this record is not cleared for redistribution. p-value p-value Quotation withheld because this record is not cleared for redistribution. standard-deviation standard-deviation about this quote: QUOTE REPAIRED: the extractor split a hyphen-space and the quote was recorded faithfully with the damage in it, which made it unanchorable. The repaired string is an EXACT SUBSTRING of the staged text - that is why it was accepted, not because a rule produced it. The text as originally recorded is kept in evidence as extracted. Quotation withheld because this record is not cleared for redistribution. t-test t-test Quotation withheld because this record is not cleared for redistribution.

The full text of this article is not redistributable, so it is not included here. The evidence sentences are withheld with it; the tag layer and its grounding status remain. The article itself is at https://arxiv.org/abs/1706.07068.