Microsoft Research's Mirage gives video generation a persistent spatial memory that doesn't forget what's around the corner

Microsoft has launched the Mirage model, which stores scene information directly in latent space instead of pixel-based point clouds. This innovation significantly reduces compute time and graphics memory while maintaining spatial consistency during long camera movements. The model enhances video generation by employing new techniques, making it more resource-efficient. However, it still struggles with reliably tracking moving objects across segments, which can affect performance accuracy in certain scenarios.