Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
194 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Generative inpainting of incomplete Euclidean distance matrices of trajectories generated by a fractional Brownian motion (2404.07029v2)

Published 10 Apr 2024 in cs.CV

Abstract: Fractional Brownian motion (fBm) features both randomness and strong scale-free correlations, challenging generative models to reproduce the intrinsic memory characterizing the underlying stochastic process. Here we examine a zoo of diffusion-based inpainting methods on a specific dataset of corrupted images, which represent incomplete Euclidean distance matrices (EDMs) of fBm at various memory exponents $H$. Our dataset implies uniqueness of the data imputation in the regime of low missing ratio, where the remaining partial graph is rigid, providing the ground truth for the inpainting. We find that the conditional diffusion generation readily reproduces the built-in correlations of fBm paths in different memory regimes (i.e., for sub-, Brownian and super-diffusion trajectories), providing a robust tool for the statistical imputation at high missing ratio. Furthermore, while diffusion models have been recently shown to memorize samples from the training database, we demonstrate that diffusion behaves qualitatively different from the database search and thus generalize rather than memorize the training dataset. As a biological application, we apply our fBm-trained diffusion model for the imputation of microscopy-derived distance matrices of chromosomal segments (FISH data) - incomplete due to experimental imperfections - and demonstrate its superiority over the standard approaches used in bioinformatics.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (43)
  1. A.Y. Alfakih. On the universal rigidity of generic bar frameworks. Contribution Disc. Math, 5(1):7–17, 2010.
  2. A. Beck and M. Teboulle. A fast iterative shrinkage-thresholding algorithm for linear inverse problems. SIAM journal on imaging sciences, 2(1):183–202, 2009.
  3. Super-resolution chromatin tracing reveals domains and cooperative interactions in single cells. Science, 362(6413):eaau1783, 2018.
  4. A. Brock and M. Hernán. Large scale gan training for high fidelity natural image synthesis. arXiv, https://arxiv.org/abs/1809.11096, 2018.
  5. Extracting training data from diffusion models. In 32nd USENIX Security Symposium (USENIX Security 23), pages 5253–5270, 2023.
  6. Iterative universal rigidity. Discrete & Computational Geometry, 53:847–877, 2015.
  7. Frameworks, tensegrities, and symmetry. Cambridge University Press, 2022.
  8. Tests for hurst effect. Biometrika, 74(1):95–101, 1987.
  9. P. Dhariwal and A. Nichol. Diffusion models beat gans on image synthesis. In Advances in Neural Information Processing Systems, pages 8780–8794, 2021.
  10. P.K. Diederik and M. Welling. Auto-encoding variational bayes. In International Conference on Learning Representations, pages 8780–8794, 2014.
  11. Euclidean distance matrices: essential theory, algorithms, and applications. IEEE Signal Processing Magazine, 32(6):12–30, 2015.
  12. Generative adversarial nets. In Advances in neural information processing systems, 2014.
  13. Characterizing the universal rigidity of generic frameworks. Discrete and Computational Geometry, 51(4):1017–1036, 2014.
  14. Positive definite completions of partial hermitian matrices. Linear Algebra and Its Applications, 58:109–124, 1984.
  15. Denoising diffusion probabilistic models. Advances in neural information processing systems, 33:6840–6851, 2020.
  16. Illuminating protein space with a programmable generative model. Nature, pages 1–9, 2023.
  17. Analyzing and improving the image quality of stylegan. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8110–8119, 2020.
  18. N. Krislock and H. Wolkowicz. Euclidean distance matrices and applications. Handbook on Semidefinite, Conic and Polynomial Optimization, page 879–914, 2012.
  19. Euclidean distance geometry and applications. SIAM Rev., 56(1):3–69, 2014.
  20. Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science, 326(5950):289–293, 2009.
  21. The geometry of graphs and some of its algorithmic applications. Combinatorica, 15(2):215–245, 1995.
  22. Repaint: Inpainting using denoising diffusion probabilistic models. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11461–11471, 2022.
  23. Fractional brownian motions, fractional noises and applications. SIAM Rev., 10(4):422–437, 1968.
  24. A. Mead. Review of the development of multidimensional scaling methods. Journal of the Royal Statistical Society: Series D (The Statistician), 41(1):27–39, 1992.
  25. K. Menger. New foundation of euclidean geometry. American Journal of Mathematics, 53(4):721–745, 1931.
  26. The random walk’s guide to anomalous diffusion: a fractional dynamics approach. Physics Reports, 339(1):1–77, 2000.
  27. Anomalous diffusion models and their properties: non-stationarity, non-ergodicity, and ageing at the centenary of single particle tracking. Physical Chemistry Chemical Physics, 16(44):24128–24164, 2014.
  28. Distance geometry: Theory, methods, and applications. New York, NY: Springer Science & Business Media, 2012.
  29. Effective hamiltonian of topologically stabilized polymer states. Soft Matter, 14(31):6561–6570, 2018.
  30. Many-body contacts in fractal polymer chains and fractional brownian trajectories. Physical Review E, 99(3):032501, 2019.
  31. Crumpled polymer with loops recapitulates key features of chromosome organization. Physical Review X, 13(4):041029, 2023.
  32. Hierarchical text-conditional image generation with clip latents. arXiv, https://arxiv.org/abs/2204.06125, 2022.
  33. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022.
  34. Photorealistic text-to-image diffusion models with deep language understanding. Advances in Neural Information Processing Systems, 35:36479–36494, 2022.
  35. J. B. Saxe. Embeddability of weighted graphs in k-space is strongly np-hard. 17th Allerton Conf. Commun. Control Comput., 1979.
  36. I.J. Schoenberg. Remarks to maurice frechet’s article“sur la definition axiomatique d’une classe d’espace distances vectoriellement applicable sur l’espace de hilbert. Annals of Mathematics, pages 724–732, 1935.
  37. Deep unsupervised learning using nonequilibrium thermodynamics. In International conference on machine learning, pages 2256–2265, 2015.
  38. Diffusion art or digital forgery? investigating data replication in diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6048–6058, 2023.
  39. Score-based generative modeling through stochastic differential equations. arXiv preprint arXiv:2011.13456, 2020.
  40. A. Vahdat and J. Kautz. Nvae: A deep hierarchical variational autoencoder. Advances in neural information processing systems, 33:19667–19679, 2020.
  41. Y. Wang and J. Cheng. Hicdiff: single-cell hi-c data denoising with diffusion models. bioRxiv, pages 2023–12, 2023.
  42. De novo design of protein structure and function with rfdiffusion. Nature, 620(7976):1089–1100, 2023.
  43. G. Young and A.S. Householder. Discussion of a set of points in terms of their mutual distances. Psychometrika, pages 19–22, 1938.

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now