Caffe is an open source deep learning framework Yangqing Jia developed during his PHD at UC Berkeley. The framework can be used for image classification, and a demo on the site lets you submit images to the system and get words back in return. I submitted an illustration of a man to the service and got back the words “consumer goods, commodity, clothing, covering, garment.” Since I don’t know a lot about machine learning or neural networks, it’s difficult for me to understand exactly what this framework is, but I just have to read more at this point. The site is comprehensive and includes links to tutorials, examples and other documentation. Parag Mital made a wrapper for this library called ofxCaffe, which he describes as follows on the GitHub page: “openFrameworks addon for visualizing and interfacing with pre-trained models in Caffe.” I’d like to try to use this library in a future project, but I have to read up first.
“NSA-Tapped Fiber Optic Cable Landing Site, Mastic Beach, New York, United States” is an interactive diptych by artist and geographer Trevor Paglen, included in the Data Issue of Dis Magazine. One interacts with the diptych using the Google Maps interface, which is a smart UI choice and an ironic gesture in light of the subject matter. The left side of the diptych features an image of Mastic Beach, one of several NSA-tapped fiber-optic cable landing sites in the US. On the right side is a collage of image and documents relating to the site — gathered from the Snowden archive and other sources — with annotations that appear when you mouse over them. The base document is a map used for marine navigation, which indicates the location of undersea cables. Paglen’s diptych avoids the abstract metaphors of mass surveillance, and instead draws from the methodologies of experimental geography. I appreciate this work’s emphasis on the physical sites and infrastructure of surveillance, and its clear presentation of multiple layers of a complex subject.