Parameter Efficient Fine-Tuning for Deep Learning-Based Full-Waveform Inversion

TL;DR

This paper introduces a task-agnostic foundational model for seismic full waveform inversion that, when fine-tuned with parameter-efficient methods, achieves high performance and efficiency across various geological scenarios.

Contribution

It presents a foundational model for FWI and demonstrates that parameter-efficient fine-tuning maintains performance while reducing computational costs.

Findings

Full fine-tuning outperforms task-specific models.

PEFT achieves comparable results to full fine-tuning.

PEFT outperforms in out-of-distribution tasks.

Abstract

Seismic full waveform inversion (FWI) has seen promising advancements through deep learning. Existing approaches typically focus on task-specific models trained and evaluated in isolation that lead to limited generalization across different geological scenarios. In this work we introduce a task-agnostic foundational model for FWI that captures general features across tasks. We first demonstrate that full fine-tuning of this foundational model outperforms task-specific models built from scratch by delivering superior performance across multiple benchmarks. Building upon this we employ parameter-efficient fine-tuning (PEFT) to further reduce computational overhead. By fine-tuning only a small fraction of the model parameters PEFT achieves comparable results to full fine-tuning while significantly lowering memory and computational requirements. Additionally, PEFT excels in out-of-distribution tasks where it outperforms both full fine-tuning and task-specific models. These findings establish the value of foundational modeling for FWI and highlight PEFT as an effective strategy for efficient and scalable adaptation across diverse tasks.