Dr Matthew Yee-King is an AI researcher, musician and leader of the BSc CS online programme at Goldsmiths, University of London. In this talk, Matthew will talk about how we can use AI techniques to map and explore spaces of creative possibilities for sound and image. He will present several examples of recent AI systems which can analyse large datasets of creative material such as music, sound and images. The examples are Music-VAE, Latent Timbre Synthesis, StyleGAN and VQGAN+CLIP. Music-VAE learns a space of symbolic music which permits interpolation between musical styles. Latent Timbre Synthesis learns a space of sonic fragments which can be explored and reconstructed into new sonic art compositions. StyleGAN provides a huge latent space of images which can be used to generate new images and to transfer semantic style elements between images. VQGAN + CLIP is a text to image system which iteratively, semantically warps a latent space to produce sequences of fascinating images. Matthew will talk about how we can work creatively with these systems to explore the space and create sounds and images that we could not otherwise create. He will present working examples of several of the systems, and provide code and resources that attendees to the seminar can use to experiment with the systems.