: A method presented at NeurIPS 2024 that uses deep text-to-image/video diffusion models to control the appearance and structure of generated media.
: A synthetic human and camera motion project presented at NeurIPS 2025 that uses deep learning to analyze camera motion and video duration. T nips.mp4
: A video dataset for deep learning exploration introduced at NeurIPS 2025 that includes over 5,000 hours of video with rich text annotations. : A method presented at NeurIPS 2024 that
If you are looking for a specific video demo of these "deep text" technologies, researchers often upload their conference presentations to the NeurIPS Virtual site as .mp4 files. T nips.mp4