Modeling dynamic neural activity by combining naturalistic video stimuli and stimulus-independent latent factors