A Data-Driven Choice of Misfit Function for FWI Using Reinforcement Learning

Abstract
In the workflow of Full-Waveform Inversion (FWI), we often tune the parameters of the inversion to help us avoid cycle skipping and obtain high resolution models. For example, typically start by using objective functions that avoid cycle skipping, and then later, we utilize the least squares misfit to admit high resolution information. Such hierarchical approaches are common in FWI, and they often depend on our manual intervention based on many factors, and of course, results depend on experience. However, with the large data size often involved in the inversion and the complexity of the process, making optimal choices is difficult even for an experienced practitioner. Thus, as an example, and within the framework of reinforcement learning, we utilize a deep-Q network (DQN) to learn an optimal policy to determine the proper timing to switch between different misfit functions. Specifically, we train the state-action value function (Q) to predict when to use the conventional L2-norm misfit function or the more advanced optimal-transport matching-filter (OTMF) misfit to mitigate the cycle-skipping and obtain high resolution, as well as improve convergence. We use a simple while demonstrative shiftedsignal inversion examples to demonstrate the basic principles of the proposed method.

Citation
Sun, B., & Alkhalifah, T. (2020). A Data-Driven Choice of Misfit Function for FWI Using Reinforcement Learning. EAGE 2020 Annual Conference & Exhibition Online. doi:10.3997/2214-4609.202010203

Publisher
EAGE Publications

Conference/Event Name
EAGE 2020 Annual Conference & Exhibition Online

DOI
10.3997/2214-4609.202010203

arXiv
2002.03154

Additional Links
https://www.earthdoc.org/content/papers/10.3997/2214-4609.202010203

Permanent link to this record

Version History

Now showing 1 - 2 of 2
VersionDateSummary
2*
2021-03-29 11:05:29
Published Confernce Paper Version
2020-02-27 05:59:27
* Selected version