Description:
 

KeyWords: #ArtificialNeuralNetwork #Backpropagation #Biomimicry #SpatialDynamics #Innovation

​

The design proposal explores a museum's spatial dynamics inspired by the Artificial Neural Network’s (ANN) structure and functioning. The museum program would include spaces for physical, digital, AR and VR AI-generated artworks and educational spaces (e.g. lecture theatre) to deliver knowledge about AI. The display galleries and the courtyard with spherical pavilions create a looped spiral circulation pattern, in which the visitors initially ascend through the galleries (feedforward) and eventually descend through 9 raised spheres (backpropagation). Additional display spaces and 10 outdoor display spheres are at the basement level. The gallery and footbridge slopes are calculated according to accessible design regulations to maximise comfort for wheelchair and stroller users.

TextToImage AI models were actively used to develop architectural design elements such as biology-based sphere finish and interior wall bas-reliefs, LED façade, and futuristic greenery at the atrium and roof terrace. The referencing and design ideation were completed with AI using the semantic method of prompt formulation, with the form following meaning. ControlNet functions were used to test different material solutions with plain conceptual massing. All artworks exhibited in the concept image were generated with AI at different stages of my academic and extra-curricular work.