PyAWD: A Library for Generating Large Synthetic Datasets of Acoustic Wave Propagation with Devito

2024-11-19

arXiv.org

AI Summary

Seismic data are inherently sparse and spatially heterogeneous, severely limiting machine learning applications in seismology; moreover, no open-source generative tools exist for large-scale, high-fidelity ground-motion simulation. To address this, we propose PyAWD—a Devito-based open-source Python library that introduces the first configurable, reproducible, and high-fidelity synthetic waveform generation framework tailored for machine learning. PyAWD supports 2D/3D heterogeneous media, fine-grained source mechanisms, and customizable acquisition parameters, and pioneers a novel data-budget analysis paradigm. Leveraging finite-difference time-domain modeling of the acoustic wave equation, PyAWD efficiently generates terabyte-scale synthetic datasets. In epicenter localization tasks, it significantly improves model accuracy under low-data regimes and enables quantitative trade-off analysis between dataset scale and predictive performance.

Technology Category

• Machine Learning: Hardware-aware ML
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Web Mining and Content Analysis: Web data generation and simulation
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Seismic data is often sparse and unevenly distributed due to the high costs and logistical challenges associated with deploying physical seismometers, limiting the application of Machine Learning (ML) in earthquake analysis. While simulation methods exist, no tool allows the generation of large datasets containing simulated measurements of the ground motion. To address this gap, we introduce PyAWD, a Python library designed to generate high-resolution synthetic datasets simulating spatio-temporal acoustic wave propagation in both two-dimensional and three-dimensional heterogeneous media. By allowing fine control over parameters such as the wave speed, external forces, spatial and temporal discretization, and media composition, PyAWD enables the creation of ML-scale datasets that capture the complexity of seismic wave behavior. We illustrate the library's potential with an epicenter retrieval task, showcasing its suitability for designing complex, accurate seismic problems that require advanced ML approaches in the absence or lack of dense real-world data. We also show the usefulness of our tool to tackle the problem of data budgeting in the framework of epicenter retrieval.

Problem

• Generates synthetic datasets for seismic wave propagation
• Addresses sparse real-world seismic data for ML applications
• Enables customizable wave simulation parameters for ML-scale datasets

Innovation

• Generates synthetic acoustic wave datasets
• Controls wave speed and media composition
• Supports 2D and 3D heterogeneous media