Experiment 07
Made with Runway
The following steps were taken every 100 steps in the entirety of the machine’s 3000 training phases. The machine was using images of chairs to learn what they look like in a progression of steps; first learning common shapes, sizes and colours. Despite a high number of training steps and images within the neural network, some of the chairs still look immensely warped.
The intention behind this experiment was to test a machine’s capacity to recognize more conventional, everyday objects and whether or not their functionality could remain intact through the decoder translation.
Made with Runway
The following steps were taken every 100 steps in the entirety of the machine’s 3000 training phases. The machine was using images of chairs to learn what they look like in a progression of steps; first learning common shapes, sizes and colours. Despite a high number of training steps and images within the neural network, some of the chairs still look immensely warped.
The intention behind this experiment was to test a machine’s capacity to recognize more conventional, everyday objects and whether or not their functionality could remain intact through the decoder translation.
