Wildfire Spread Scenarios: Increasing Sample Diversity of Segmentation Diffusion Models with Training-Free Methods

2026-03-20

AI Summary

This work addresses the challenge of generating diverse and plausible multimodal predictions in uncertain environments—such as wildfire spread—where conventional diffusion models suffer from low sampling efficiency. The authors propose a training-free sampling strategy that, for the first time, adapts particle guidance and diversity-enhancing techniques like SPELL from image generation to discrete segmentation tasks. A lightweight clustering-based approach is further introduced to boost sample diversity. To support research in this domain, the authors construct and publicly release MMFire, a novel wildfire simulation dataset that fills a critical gap in existing benchmarks. Experiments on both MMFire and Cityscapes demonstrate significant improvements in HM IoU* by 7.5% and 16.4%, respectively, achieving markedly enhanced prediction diversity with negligible computational overhead and no degradation in output quality.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Web data generation and simulation
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Predicting future states in uncertain environments, such as wildfire spread, medical diagnosis, or autonomous driving, requires models that can consider multiple plausible outcomes. While diffusion models can effectively learn such multi-modal distributions, naively sampling from these models is computationally inefficient, potentially requiring hundreds of samples to find low-probability modes that may still be operationally relevant. In this work, we address the challenge of sample-efficient ambiguous segmentation by evaluating several training-free sampling methods that encourage diverse predictions. We adapt two techniques, particle guidance and SPELL, originally designed for the generation of diverse natural images, to discrete segmentation tasks, and additionally propose a simple clustering-based technique. We validate these approaches on the LIDC medical dataset, a modified version of the Cityscapes dataset, and MMFire, a new simulation-based wildfire spread dataset introduced in this paper. Compared to naive sampling, these approaches increase the HM IoU* metric by up to 7.5% on MMFire and 16.4% on Cityscapes, demonstrating that training-free methods can be used to efficiently increase the sample diversity of segmentation diffusion models with little cost to image quality and runtime. Code and dataset: https://github.com/SebastianGer/wildfire-spread-scenarios

Problem

• wildfire spread
• segmentation diffusion models
• sample diversity
• ambiguous segmentation
• multi-modal prediction

Innovation

• diffusion models
• sample diversity
• training-free sampling
• ambiguous segmentation
• wildfire spread prediction