As a part of our research project "What is the wind thinking about?" for the Shishim Hill artist residence, we planned to measure wind speed in the Shishim Hill area using video recordings of the fluctuations of fabric flags tied to ropes. We wanted to study the data by passing it through an artificial neural network, which could catch a logic in the way the wind blows, and then translate it into a pattern on the LED screen. While working on the project, we assumed that the neural network understands the wind, and it can predict changes in the graphs of its speed. Therefore, we created a recurrent neural network that solves this problem, and the values of the internal memory cells of a working neural network were shown as light strips demonstrating 'thoughts of the wind'.
The video with the light strips was shown at the final event at the residence as a 3D model of the LED installation, with its colours changing depending on the wind speed in the Shishim Hill area in real time. We then presented the project at the UNDARK festival as a finished installation demonstrating the recorded video.