vellum

[doubleedgedhead]

[lostfunzones]

via: equatorjournal

[vellum]

This episode of guitar moves makes me so happy.

[vellum]

“Game Theory Beijing,” by Experimental Jetset

[parallelpractice]

Text-flow.

[prostheticknowledge]

LipNet

Deep Learning research from the University of Oxford and Google DeepMind can accurately deduce sentences from visual analysis of speaking:

Lipreading is the task of decoding text from the movement of a speaker’s mouth. Traditional approaches separated the problem into two stages: designing or learning visual features, and prediction. More recent deep lipreading approaches are end-to-end trainable (Wand et al., 2016; Chung & Zisserman, 2016a). All existing works, however, perform only word classification, not sentence-level sequence prediction. Studies have shown that human lipreading performance increases for longer words (Easton & Basala, 1982), indicating the importance of features capturing temporal context in an ambiguous communication channel. Motivated by this observation, we present LipNet, a model that maps a variable-length sequence of video frames to text, making use of spatiotemporal convolutions, an LSTM recurrent network, and the connectionist temporal classification loss, trained entirely end-to-end. To the best of our knowledge, LipNet is the first lipreading model to operate at sentence-level, using a single end-to-end speaker-independent deep model to simultaneously learn spatiotemporal visual features and a sequence model. On the GRID corpus, LipNet achieves 93.4% accuracy, outperforming experienced human lipreaders and the previous 79.6% state-of-the-art accuracy. 

More Here

[prostheticknowledge]

SEMI SENSELESS DRAWING MODULES

Installations by Yamaguchi Takahiro and So Kanno explores themes related to imitation and replication through mark making using robotics, computer vision and machine learning.

The first project, Replicate, is a workshop featuring a group of schoolchildren drawing on a wall, whose movements are captured using depth-camera data:

The second project, Letters, uses handwriting written in nearby book, breaking down the penstrokes into individual parts with machine learning, and reconstructs them as markings for the drawing robot to replicate:

SDM2 - Letters, which theme is learning, generates lines looks like letters but doesn’t make sense by machine learning system which learnd shapes and patterns of hand writing strokes with ignoring meaning of them. On street in some international city, you hear a lot of different kind of language, and it’s possible to distinguish what language it is without understanding its meaning. It means human learns sound before meaning. This project attempts to make a same phenomenon with artificial intelligence and hand writing letters. Lines generated by system with biased learning and removing meaning from hand writing letters, how does it look?

The installations are currently on view at the 21_21 Design Sight Gallery, Roppongi, Tokyo, which you can find out more here

[prostheticknowledge]

Tilt-Ikeda

Creative coder Patricio Gonzalez Vivo has been developing many graphical shaders for maps recently, and his latest is a tribute to well-renowned tech artist Ryoji Ikeda.

Try it out for yourself here

[vellum]

Tamas Kemenczy unpacks the set design and storytelling techniques in KRZ

[vellum]

Jake Elliott on experimental cinema and its influences on KRZ

[vellum]

(via How Two Ex-Amazon Engineers Invented a New Type of Customizable Force Interface )

[vellum]

JAWS: The Server-less Application Stack – Uses bleeding-edge AWS services to redefine how to build massively scalable (and cheap) apps!

[vellum]

(via Cheeseburgers, Weed & The Seahawks: Chef's Night Out in Seattle with Josh Henderson - YouTube)

[vellum]

(via Bill's Design Talks: Pentagram - YouTube)

[archivesofamericanart]

In the not-too-distant past, contacts were stored not in pocket-sized computers but in pocket-sized address books. Our new exhibit gives you a glimpse into the grimy, dog-eared pages of well-known artists’ “little black books.” From Ad Reinhardt’s address book that fits in the palm of your hand, to Holly Solomon’s hulking double rotary Rolodex, these documents help us understand the social networks in which these artists lived and worked.

Little Black Books: Address Books from the Archives of American Art is on view in the Lawrence A. Fleischman Gallery from August 7 - November 1, 2015. 

gif from: Ad Reinhardt’s address book, circa 1950-1965. Ad Reinhardt papers, Archives of American Art, Smithsonian Institution.

[brucesterling]

*Paper.  And they, like, mailed stuff.  About “art.”

[vellum]

(via Google I/O 2015 - A little badass. Beautiful. Tech and human. Work and love. ATAP. - YouTube)

The state of the art in touch technologies seems good.

[vellum]

(via Stanford engineers build a water-droplet based computer that runs like clockwork - YouTube)