top of page
Lesson
HD Video (5:40), Monitor, 2021


石橋 友也  新倉 健人
    Tomoya Ishibashi Kento Niikura  

Photo: Ken Kato Photo courtesy: Tokyo Arts and Space

左のモニターには有名絵画、中央のモニターには「画像の説明文を生成するAI」が生成した絵画の説明文、右側のモニターには「テキストから画像を生成するAI」が中央のテキストを元に生成した画像を表示する。謂わば「AIが解釈した絵画」を生成する一連のプロセスを提示した。

ミレーの《落ち穂拾い》を農婦の腕を誤認識し「象の群れ」と解釈する等、現状のAIの精度故のズレがしばしば発生している。絵画はジョットから始まり、順に20世紀絵画へと進んでいく。AIは写実的な絵画では比較的妥当な解釈をするが、抽象画へと進むうちにその乖離は大きくなる。一方で、カンディンスキーを「ハサミの束」と解釈した上で、色彩や構図の奇妙な一致が見られるイメージを生成したり、ポロックを「森の中の木」とするような発見性のある解釈を返してきたりもする。

タイトルである《Lesson》は人間の美術鑑賞や作品制作の成長過程になぞらえたもの。本作は機械の視点を通して、芸術作品の解釈や創造行為の主体性や偶然性を問う試みである。

The left monitor shows a famous painting, the center monitor shows a description of the painting generated by the image description generator, and the right monitor shows an image generated by the text-to-image generator based on the text in the center. In other words, a series of processes to generate AI-interpreted paintings were presented.
 

Due to the accuracy of current AIs, mismatches can often happen, such as misinterpreting a farmer's arm in Millet's The Gleaners as a "herd of elephants." Starting with Giotto and moving on to 20th-century paintings, the AI makes relatively reasonable interpretations of realistic paintings, but the gap widens as it moves on to abstract paintings. In addition to interpreting Kandinsky as "a bunch of scissors," it can also generate images with strange coincidences of color and composition and return heuristic interpretations including Pollock as "a tree in the forest." 
 

The title "Lesson" is a metaphor for the process of human growth in art appreciation and creation. This work is an attempt to question the subjectivity and contingency of the interpretation of artworks and the act of creation through the perspective of a machine.
 

s_KNK1398.jpg
s_KNK1375.jpg

コトバノキカイ/Machine of Words, Tokyo Arts and Space hongo (6/1/2021-6/20/2021)

​引用作品

The Crucifixion of Our Lord Christ, Giotto di Bondone, ca.1300
The Last Supper, Leonardo da Vinci, ca.1495
Aldobrandini Madonna, Raffaello Santi, ca.1500
The Milkmaid, Johannes Vermeer, ca. 1658
The Gleaners, Jean-François Millet, 1857
The Luncheon on the Grass, Édouard Manet, 1863
Sunflowers (F456), third version: blue green background, Vincent van Gogh, 1888
The Young Ladies of Avignon, Pablo Ruiz Picasso, 1907
Composition VII, Wassily Kandinsky, 1913
The Treachery of Images, René Magritte, 1929
Composition with Red, Blue and Yellow, Piet Mondrian, 1930
Mural on Indian Red Ground, Jackson Pollock, 1950
Just what is it that makes today's homes so different, so appealing?, Richard Hamilton, 1956
COMPOSITION, Mark Rothko, 1959
Concetto Spaziale/Attese, Lucio Fontana, 1961
Three Studies for Portrait of George Dyer (on pink ground), Francis Bacon, 1964
Anna's Light, Barnett Newman, 1968

引用コード

CNN+LSTM https://github.com/tensorflow/docs-l10n/blob/master/site/ja/tutorials/text/image_captioning.ipynb

L-XMERT https://vision-explorer.allenai.org/text_to_image_generation

bottom of page