Research map / Image generation / Autoregressive and latent generation
Hamiltonian Latent Operators for content and motion disentanglement in image sequences
Asif Khan, Amos Storkey
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use HALO instead of DSVAE, MoCoGAN, or S3VAE | MUG · MUG | Accuracy: 0.543; 0.631; 0.705; HALO: 0.929 | Improved |
| Use HALO without action labels instead of the baselines | MUG · MUG | Accuracy: 0.543; 0.631; 0.705; HALO: 0.750 | Improved |
| Use HALO instead of the baselines | Sprites · Sprites | Accuracy: 0.907; 0.928; 0.994; HALO: 1.000 | Mixed |
| Use HALO instead of the baselines | MUG · MUG | IS: 3.607; 4.655; 5.078; HALO: 5.312 | Improved |
| Use HALO instead of the baselines | Sprites · Sprites | IS: 8.331; 8.182; 8.636; HALO: 7.374 | Hurt |
| Restrict the Hamiltonian operator to a skew-symmetric form (ablation) | HALO · Sprites | SSIM: 0.982±0.005 -> 0.950±0.021 | Hurt |
| Restrict the Hamiltonian operator to a skew-symmetric form (ablation) | HALO · MUG | SSIM: 0.797±0.003 -> 0.791±0.003 | Hurt |
| Restrict the Hamiltonian operator to a skew-symmetric form (ablation) | HALO · Sprites | PSNR: 36.76±1.096 -> 33.88±2.03 | Hurt |
| Restrict the Hamiltonian operator to a skew-symmetric form (ablation) | HALO · MUG | PSNR: 24.49±0.099 -> 24.25±0.094 | Hurt |
| Restrict the Hamiltonian operator to a skew-symmetric form (ablation) | HALO · Sprites | MSE: 0.0005±0.0002 -> 0.0026±0.0012 | Hurt |
| Restrict the Hamiltonian operator to a skew-symmetric form (ablation) | HALO · MUG | MSE: 0.0040±0.0001 -> 0.0044±0.0001 | Hurt |
| Replace Hamiltonian dynamics with a linear dynamical model (ablation) | HALO · Image-to-sequence and motion transfer | controlled generation and disentanglement | Hurt |
Closest papers by meaning
- Continuity-Preserving Convolutional Autoencoders for Learning Continuous Latent Dynamical Models from Images (ICLR 2025 · similarity 0.84)
- Exploring Invariance in Images through One-way Wave Equations (ICML 2025 · similarity 0.84)
- LaM-SLidE: Latent Space Modeling of Spatial Dynamical Systems via Linked Entities (NeurIPS 2025 · similarity 0.84)
- Generative Image Dynamics (CVPR 2024 · similarity 0.84)
- Isometric Representation Learning for Disentangled Latent Space of Diffusion Models (ICML 2024 · similarity 0.84)
- Flow Matching with General Discrete Paths: A Kinetic-Optimal Perspective (ICLR 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
