43733MediaFinder: a content-based recommendation system for user-generated audio-visual contentClosedStudyInnovate UKIn recent years, the use of deep learning methods has pushed the boundaries of AI, leading to immense advances in a number of cutting-edge technologies, including self-driving vehicles, facial recognition systems and fraud detection services. At Mashtraxx Ltd (MXX), a team of experienced AI researchers has brought the power of deep learning to music recommendation. The science is inspired by cutting-edge computer vision technologies, which successfully solved complex tasks such as detecting objects in satellite pictures or recognizing individuals in images. MXX has developed a suite of tools for intelligent music production, editing and consumption. Among our new research is a content-based **music recommendation system,** which allows users to search large music collections using an audio track as a query. Through initial discussions of this technology with social media platforms, and our own mobile app, we have identified an unmet need for a complementary **video recommendation system** for online platforms, which can cope with the more difficult challenges centred around user-generated videos as well as music. This MediaFinder proposal will be a substantial step-change in recommendation technology, which will disrupt several sectors.