Aaron here today with an Art from Guantánamo Bay Update; Comic Book Heroes of Taiwan; Austin English on the End of Krazy Kat (and Superman Comics); More Machine Learning


Ode to the Sea
An update regarding the Art from Guantánamo Bay post from a few weeks back, via at the Comic Book Legal Defense Fund:

Fortnunately, the Pentagon is backing off its previous threat of incinerating art by the Guantánamo Bay detainees. Unfortunately, they still maintain all of it is the property of the U.S. military and will not be permitted to leave the prison.

The exhibit is free and open to the public and runs through January 26, 2018, at Haren Hall (899 10th Avenue at 59th Street, on the 6th floor in the President’s Gallery, New York City).  The exhibit is open Monday to Friday, from 9:00 a.m. to 5:00 p.m. (except January 15).


Taiwan’s comic book heroes
Via the BBC:

Hsu Mao-sung, one of Taiwan’s earliest comic book artists, says it may be too late to save the industry.

His generation endured censorship during the martial law period from the 1960s to 80s, but he says it didn’t kill off his passion for drawing.


‘Sequence from June 4th 1944 Krazy Kat strip by George Herriman’

Anything But Reality
At TCJ, Austin English takes a long look at some comics from the 1940s-1960s:

The strip that follows is June 11th, with the stick now re-purposed to prop up the scarecrow. So, the final interactions of Krazy Kat and Ignatz map out like this: Ignatz tricks Krazy into getting hit by a brick toppling from a growing tree, Ignatz passingly confuses Krazy about oak trees vs. olive trees, Ignatz works to transform a weasel’s color in an issue whose subtleties are (seemingly) obscure to Krazy and finally Ignatz’ brick is discovered with a stick from Offisa Pup while Ignatz writhes in agony, his body beneath ground. After these events, the scarecrow emerges and Ignatz discards his brick. For a brief moment the strip is silent and empty.

The tone of the line defined itself in 1958, partly by a constraint that the comics world imposed on itself. Physical violence, while not banned by the comics code, was to be avoided. So Weisinger took a line of superhero comics and instead of emphasizing brawn, made them into intricate puzzles. Now, these are not puzzles that use the unreliable world as one of its factors in the way a Raymond Chandler novel might. Instead, every story has no relation to anything real (except basic outlines of things like ‘I have a job at a newspaper’ or ‘humans need to eat food to survive’ or ‘ice is cold’)—the comic book world of Superman itself is the only thing the stories use to set up their questions and render solutions. These are mysteries whose only logic is cartooning, and I’d argue that they are more beautiful as science fiction than anything EC ever published.

Panel from ‘The Shrinking Superman’ by Otto Binder (script) and Wayne Boring (art), Action Comics #245, 1958


Figure 1: A real-world attack on VGG16, using a physical patch generated by the white-box ensemble
method described in Section 3. When a photo of a tabletop with a banana and a notebook (top
photograph) is passed through VGG16, the network reports class ’banana’ with 97% confidence (top
plot). If we physically place a sticker targeted to the class “toaster” on the table (bottom photograph),
the photograph is classified as a toaster with 99% confidence (bottom plot). See the following video
for a full demonstration: https://youtu.be/i1sp4X57TL4

Adversarial Patch
An update from the machine learning front, via Boing Boing:

Another key difference here: the researchers achieve their best results using a “white box” technique where they get to design their patches through detailed knowledge of the AI they’re targeting, unlike other adversarial preturbations, which achieved good results with “black box” constraints (designing attacks without any technical knowledge of the AI). The patches they created didn’t work very well on other AI image classifiers.

We show that we can generate a universal, robust, targeted patch that fools classifiers regardless of the scale or location of the patch, and does not require knowledge of the other items in the scene that it is attacking. Our attack works in the real world, and can be disguised as an innocuous sticker. These results demonstrate an attack that could be created offline, and then broadly shared. There has been substantial work on defending against small perturbations to natural images, at least partially motivated by security concerns Part of the motivation of this work is that potential malicious attackers may not be concerned with generating small or imperceptible perturbations to a natural image, but may instead opt for larger more effective but noticeable perturbations to the input – especially if a model has been designed to resist small perturbations.

Many ML models operate without human validation of every input and thus malicious attackers may not be concerned with the imperceptibility of their attacks. Even if humans are able to notice these patches, they may not understand the intent of the patch and instead view it as a form of art. This work shows that focusing only on defending against small perturbations is insufficient, as large, local perturbations can also break classifiers.


Frank Santoro and Simon Hanselmann, CAB 2013 – photo by Chris Anthony Diaz, colored by Graham Willcox

The Winter Semester of thee Santoro Correspondence Course for Comic Book Makers starts January 16th 2018! 8 weeks – 500 bux – coaching for as long as you need. The course is hard, but Frank will push your comics making practice to a new level, getting you to think about timing and color in new ways. His experience and ideas have influenced the likes of Connor Willumsen, Michael DeForge, and Simon Hanselmann (quote “I consider Frank Santoro to be my L. Ron Hubbard”) among many others. Dig into something new in the new year!

Full details and how to apply can be found HERE.


Vision Box – 1-9-2018 – by Cameron Arthur

Share this page:

Leave a Reply

Your email address will not be published. Required fields are marked *