Generative diffusion model surrogates for mechanistic agent-based biological models

2025/05/01 by Tien Comlekoglu, Comlekoglu, Tien, J. Quetzalcoatl Toledo-Marín +10 · 2 voices · 1 citation
Biochemistry, Genetics and Molecular Biology · Computer Science · Mathematics · #Computer Vision and Pattern Recognition (cs.CV) #Emerging Technologies (cs.ET) #FOS: Biological sciences #FOS: Computer and information sciences #Mathematical Biology Tumor Growth #Performance (cs.PF) #Quantitative Methods (q-bio.QM) #cs.CV #cs.ET #cs.PF #q-bio.QM

paper · pdf · doi:10.48550/arxiv.2505.09630

openalex publication_date 2025/05/01 · arxiv published 2025/05/01 · openalex created_date 2025/10/10 · arxiv updated 2025/10/23 · openalex updated_date 2026/07/28

Abstract

Mechanistic, multicellular, agent-based models are commonly used to investigate tissue, organ, and organism-scale biology at single-cell resolution. The Cellular-Potts Model (CPM) is a powerful and popular framework for developing and interrogating these models. CPMs become computationally expensive at large space- and time- scales making application and investigation of developed models difficult. Surrogate models may allow for the accelerated evaluation of CPMs of complex biological systems. However, the stochastic nature of these models means each set of parameters may give rise to different model configurations, complicating surrogate model development. In this work, we leverage denoising diffusion probabilistic models to train a generative AI surrogate of a CPM used to investigate in vitro vasculogenesis. We describe the use of an image classifier to learn the characteristics that define unique areas of a 2-dimensional parameter space. We then apply this classifier to aid in surrogate model selection and verification. Our CPM model surrogate generates model configurations 20,000 timesteps ahead of a reference configuration and demonstrates approximately a 22x reduction in computational time as compared to native code execution. Our work represents a step towards the implementation of DDPMs to develop digital twins of stochastic biological systems.

Citations

• Surrogate modeling of Cellular-Potts Agent-Based Models as a segmentation task using the U-Net neural network architecture
• Zephyr quantum-assisted hierarchical Calo4pQVAE for particle-calorimeter interactions
• CaloChallenge 2022: A Community Challenge for Fast Calorimeter Simulation
• Avoiding mode collapse in diffusion models fine-tuned with reinforcement learning
• Conditional Image Synthesis with Diffusion Models: A Survey
• Modeling the roles of cohesotaxis, cell-intercalation, and tissue geometry in collective cell migration of Xenopus mesendoderm
• A Comprehensive Evaluation of Generative Models in Calorimeter Shower Simulation
• Guiding a Diffusion Model with a Bad Version of Itself
• DGRM: Diffusion-GAN recommendation model to alleviate the mode collapse problem in sparse environments
• DiffDA: a Diffusion Model for Weather-scale Data Assimilation
• A parallelized cellular Potts model that enables simulations at tissue scale
• Analyzing and Improving the Training Dynamics of Diffusion Models
• Residual Corrective Diffusion Modeling for Km-scale Atmospheric Downscaling
• Interpolating between Images with Diffusion Models
• Adding Conditional Control to Text-to-Image Diffusion Models
• Adding Conditional Control to Text-to-Image Diffusion Models
• Analyzing the Performance of Deep Encoder-Decoder Networks as Surrogates for a Diffusion Equation
• Classifier-Free Diffusion Guidance
• Elucidating the Design Space of Diffusion-Based Generative Models
• Diffusion Models Beat GANs on Image Synthesis
• EfficientNetV2: Smaller Models and Faster Training
• Deep learning approaches to surrogates for solving the diffusion equation for mechanistic real-world simulations
• Combating Mode Collapse in GAN training: An Empirical Analysis using\n Hessian Eigenvalues
• Denoising Diffusion Probabilistic Models
• Learning Everywhere: A Taxonomy for the Integration of Machine Learning and Simulations
• Deep Learning the Physics of Transport Phenomena
• Comparing individual-based approaches to modelling the self-organization of multicellular tissues
• ImageNet Large Scale Visual Recognition Challenge
• ImageNet Large Scale Visual Recognition Challenge

Cited by

• Surrogate modeling of Cellular-Potts Agent-Based Models as a segmentation task using the U-Net neural network architecture

Discussions

• My dad's research is using diffusion models as fast surrogates for slow, stochastic agent‑based biology simulations. https://arxiv.org/pdf/2505.09630 Always interesting to learn different uses of gene [bsky, 4 points, 0 comments]
• My dad's research is using diffusion models as fast surrogates for slow, stochastic agent‑based biology simulations. https://arxiv.org/pdf/2505.09630 Always interesting to learn different uses of gene [bsky, 0 points, 0 comments]

Related

• Surrogate modeling of Cellular-Potts Agent-Based Models as a segmentation task using the U-Net neural network architecture
• Revealing Subtle Phenotypes in Small Microscopy Datasets Using Latent Diffusion Models
• An Organism Starts with a Single Pix-Cell: A Neural Cellular Diffusion for High-Resolution Image Synthesis
• Towards Learned Simulators for Cell Migration
• Denoising Diffusion Probabilistic Models for Generation of Realistic Fully-Annotated Microscopy Image Data Sets