coming project

0 title [Exhibition]COMO&HAPPY SCREEN_MAY 2016《Learning About》 date 2016.04.28
file hit 11,825

COMO&HAPPY SCREEN_MAY 2016 《Learning About》
2016. 05. 02 – 06. 30

Exhibition venue & Open Hours
 - COMO (SKT-Tower, Euljiro 1-ga Exit no.4 / Daejeon SKT Dunsan Bldg) 
  - Happy Screen (SK Bldg B1/4F)
   11:30am - 01:00pm / 02:30pm - 04:00pm
 * closed on weekends and public holidays 
Entrance Fee Free
Exhibition Organizer Art Center Nabi, SK Telecom, SK innovation
Exhibition Artist Sebastian Schmieg
Exhibition Curator SeongEun An
Main Curator PilUng Kang
Video Editing JiHye Gong
Video Transmission YoungHo Lee, YeongHwan Kim
Design YunJeong Kim
Exhibition Inquiry  02-2121-0933

What differentiates the future of Artificial Intelligence from other existing technologies, which makes it more promising yet scary, is the aspect of learning that it learns from information it is given. While other existing technologies receive information as an input, Artificial Intelligence shows an incredible pace of development as it teaches itself by learning from big data.

Then, how artificial intelligent ‘learn’ by itself?

The self-learning of artificial intelligence often shows better results than humans depending on which data it learns from. However, how it processes the learning process is still unknown or requires human labor. Thus, there are different speculations about how people’s lives will change by the development of this technology and how artificial intelligence teaches itself and learns by itself.

The current edition of COMO&HAPPY SCREEN presents an exhibition 《Learning About》. The exhibition revolves around the issues of the learning method of artificial intelligence, continuing the theme of the previous exhibition that introduced how artificial intelligence recognize the world by showing the difference in human perspective and computer vision.

Learning About presents Sebastian Schmieg, a media artist based in Berlin. Schmieg has been producing works based on the web, appropriating technologies such as machine learning, computer vision, and network. The current exhibition presents 〈LSTM(Long Short Term Memory〉(2015) and 〈Segmentation.Network〉(2016) that approach how humans perceive the world and adapt to it with the development of new technologies.

By forecasting the future by borrowing words by a futurologist and revealing the hidden human labor sacrificed for enhancing the cognitive accuracy of artificial intelligence, Sebastian Schmieg’s work raises questions about the way artificial intelligence teaches itself and learns by itself while also examining the possibility of machine learning and its manual and labor-intensive aspects.

As the current period is witnessing the change of technological paradigm by the innovative development of high technologies such as smart mobile devices and artificial intelligence, it is not easy to overview tasks given to our perceptional organs only by seeing and observing them. The tasks of perception in this paradigm shift will be overcome by its tactile reception or habituation.*

We are already living with ‘domesticized’ services that employ artificial intelligence without recognizing them. It is hoped that the current exhibition will provide a ground for realizing the hidden face of artificial intelligence that has been considered only in terms of the mechanism of technological development as well as an occasion to discuss different ways of considering the ways through which artificial intelligence teaches and learns by itself.

* Quoted from Walter Benjamin’s The Work of Art in the Age of Mechanical Reproduction. 

About the works

LSTM(Long Short Term Memory), 2015
a result of the Goethe@WRO residency

〈LSTM(Long Short Term Memory)〉(2015) is a work where an artificial neural network learns the whole books written by Ray Kurzweil, a futurologist and the director of engineering at Google, and produces new informational texts based on the knowledge learned from Kurzweil’s books. The work employs the technology of deep learning, which makes the level of learning differently in each day and leads to more elaborate texts as it continues its learning process. What is interesting in this work is the fact that an artificial intelligence, an issue that is at the epicenter of controversy with regards to its impact on the future of human race, tells about the future by learning books by a representative futurologist Ray Kurzweil that deal with issues such as the technological singularity, accelerated development of artificial intelligence, thoughts on the future, and medical revolution. The work leads the viewers to reflect on the level of learning that artificial intelligence possesses and its potential of development by presenting a new way of producing texts.


Segmentation.Network, 2016

Different from 〈LSTM〉 that appropriates an artificial neural network, 〈Segmentation.Network〉(2016) uses human labor, which is dataset manually created by crowd workers for enhancing the accuracy of object recognition of artificial intelligence. The work automatically displays drawings that are based on the dataset. The recognition rate of artificial intelligence is most accurate when the outlines of objects are provided in the form of two-dimensional images. Thus, corporations have been producing a huge amount of image dataset by temporary workers to enhance the accuracy when creating artificial intelligence programs for object recognition. The dataset used in 〈Segmentation.Network〉 is the one used by Microsoft’s COCO(Common Objects in Context) where crowd workers created dataset of approximately six hundred thousand images on Flickr website. The work traces the data used to enhance the accuracy of artificial intelligence and focuses on the hidden human labor behind machine learning, leading to a reflection of the way artificial intelligence learns information.

〈LSTM〉(2016), 53,000 x 1,000mm, LED panel, Neural Network Website, Teleprompter

〈LSTM〉(2016), 53,000 x 1,000mm, LED panel, Neural Network Website, Teleprompter

〈Segmentation.Network〉(2016), 5,300 x 3,200mm, LED panel, Website

〈LSTM〉(2015), Ceiling 30,300 x 1,280mm, Pillar(4ea) 1,024 x 6,144mm, LED panel, Neural Network Website, Teleprompter

〈LSTM〉(2015), Ceiling 30,300 x 1,280mm, Pillar(4ea) 1,024 x 6,144mm, LED panel, Neural Network Website, Teleprompter


Sebastian Schmieg

Sebastian Schmieg(b. 1983) is a media artist based in Berlin. Schmieg has been producing web-based works that appropriate technologies such as machine learning, computer vision, and network. With an interest in the use and exploration of media that flexibly connect human and technology, Schmieg has been experimenting with the point of connection between the two. Schmieg participated in Solve for X, a think tank project by Google to incite collaboration to solve global issues, and how to appear offline forever(2015), a collaboration with a British media art institution AND(Abandon Normal Devices). Schmieg is also actively participating a number of exhibitions such transmediale in Germany.

Selected Exhibitions
2016 setBackground, EMBASSY Gallery, Edinburgh, United Kingdom
2015  LSTM (Long Short Term Memory), Online, Cointemporary 
2015 The ROCI Road to Peace: Experiments in the Unfamiliar, Academy Art Museum, Easton, USA
2015 Phaenomenale, phæno, Wolfsburg, Germany
2015 Abandon Normal Devices, Grizedale Forest, UK
2015 ISEA, Vancouver, Canada
2015 NEWMAN, Druskininkai, Lithuania
2015 Offline Art: Are you still there?, Museum Angewandte Kunst, Frankfurt, Germany
2015 16th media art biennale WRO 2015, Wroclaw, Poland
2015 The New Black (Resonate), Gallery 12 New Media HUB, Belgrade, Serbia
2015 Transmediale: Capture All, Berlin, Germany
2014 Parallel Universes, Online/Screensaver, ScreenSaverGallery 
2014 Transmediale: Afterglow, Haus der Kulturen der Welt, Berlin, Germany 
2013 Art Hack Day Berlin: Going Dark, LEAP, Berlin, Germany
2013 Vanishing Point, bitforms gallery, New York, USA
2012 reflecting on networks, km temporaer, Berlin, Germany
2011 lab.30, Augsburg, Germany
2010 FILE(Electronic Language International Festival), Sao Paulo, Brazil
2009 A Maze, Berlin, Germany
back to list