Generative Adversarial Neural Cellular Automata
This work addresses the adaptability issue in generative models for researchers in machine learning and computer vision, though it is incremental as it builds on existing Neural Cellular Automata and GAN concepts.
The paper tackles the limitation of Neural Cellular Automata requiring separate models for each image generation scenario by introducing a single model that uses different initial environments to produce multiple outputs, and it proposes GANCA, which combines Neural Cellular Automata with Generative Adversarial Networks for adversarial training. The results show that a single model can learn several images from varied inputs, and the adversarially trained model improves drastically on out-of-distribution data compared to supervised training.
Motivated by the interaction between cells, the recently introduced concept of Neural Cellular Automata shows promising results in a variety of tasks. So far, this concept was mostly used to generate images for a single scenario. As each scenario requires a new model, this type of generation seems contradictory to the adaptability of cells in nature. To address this contradiction, we introduce a concept using different initial environments as input while using a single Neural Cellular Automata to produce several outputs. Additionally, we introduce GANCA, a novel algorithm that combines Neural Cellular Automata with Generative Adversarial Networks, allowing for more generalization through adversarial training. The experiments show that a single model is capable of learning several images when presented with different inputs, and that the adversarially trained model improves drastically on out-of-distribution data compared to a supervised trained model.