\n
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F22830928%2Fc0e61fe7cdf87acd5a02b112313f31cb%2F1.png?generation=1749086325130384&alt=media\")
\n
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F22830928%2Fd4df18a0e1f4151d15bdb66b7aeebca4%2F2.png?generation=1749086341724357&alt=media\")
\n![]
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F22830928%2F3c972cbb423c8fe24131507b77f7f1e8%2F3.png?generation=1749086370109605&alt=media\")
can we get some idea from this paper? https://ieeexplore.ieee.org/document/10697189
\n
\n
\n![]
can we get some idea from this paper? https://ieeexplore.ieee.org/document/10697189
\n
\n
\n![]
I did some googling and followed some links to try to understand what \"Q\" is all about… I think it is a measure of attenuation of the wave as it travels through a medium; this is in addition to spherical spreading and reflections. Here's one basic description of it (from Hermana et al. 2012):
\n\n\n\"Attenuation is usually addressed as quality factor (Q)
\n
\n term which is always associated with its intrinsic properties
\n of medium: pore fluid content and lithology variation.
\n Medium with high attenuation such as gas bearing strata
\n has low quality factor. In this condition, energy of seismic
\n wave will be absorbed strongly by the medium. Attenuation
\n is more sensitive than velocity due to fluid content changes.\"
Also, attenuation is usually frequency dependent (i.e., dispersion) causing the wave to change shape and spectral characteristics. So, Q methods may not be applicable because attenuation and dispersion are not included in the simulations used in this competition.
\nIt may be possible to adapt their general NN+Physics scheme/architecture as a way to include physics that is appropriate to this GWI situation, but that would take a bigger brain than mine 😄
", "rawMarkdown": "I did some googling and followed some links to try to understand what \"Q\" is all about... I think it is a measure of **attenuation of the wave** as it travels through a medium; this is in addition to spherical spreading and reflections. Here's one basic description of it (from [Hermana et al. 2012](https://www.ukm.my/jsm/pdf_files/SM-PDF-41-8-2012/03%20Maman%20Hermana.pdf)):\n\n>\"Attenuation is usually addressed as quality factor (Q)\nterm which is always associated with its intrinsic properties\nof medium: pore fluid content and lithology variation.\nMedium with high attenuation such as gas bearing strata\nhas low quality factor. In this condition, energy of seismic\nwave will be absorbed strongly by the medium. Attenuation\nis more sensitive than velocity due to fluid content changes.\"\n\nAlso, attenuation is usually frequency dependent (i.e., dispersion) causing the wave to change shape and spectral characteristics. **So, Q methods may not be applicable because attenuation and dispersion are not included in the simulations used in this competition.**\n\nIt may be possible to adapt their general NN+Physics scheme/architecture as a way to include physics that is appropriate to this GWI situation, but that would take a bigger brain than mine 😄", "votes": null }, { "id": "3217377", "postDate": "06/05/2025 00:55:11", "content": "yeah. I always try to build a neural networks by the physics-based.
", "rawMarkdown": "yeah. I always try to build a neural networks by the physics-based.", "votes": null }, { "id": "3217388", "postDate": "06/05/2025 01:17:02", "content": "![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F22830928%2F36592486e50139ff97c7468327f1abe8%2F1.png?generation=1749086215568973&alt=media\")
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F22830928%2F44f9a6d25322a318aa103ea9584ab278%2F2.png?generation=1749086252674721&alt=media\")
I did some googling and followed some links to try to understand what \"Q\" is all about… I think it is a measure of attenuation of the wave as it travels through a medium; this is in addition to spherical spreading and reflections. Here's one basic description of it (from Hermana et al. 2012):
\n\n\n\"Attenuation is usually addressed as quality factor (Q)
\n
\n term which is always associated with its intrinsic properties
\n of medium: pore fluid content and lithology variation.
\n Medium with high attenuation such as gas bearing strata
\n has low quality factor. In this condition, energy of seismic
\n wave will be absorbed strongly by the medium. Attenuation
\n is more sensitive than velocity due to fluid content changes.\"
Also, attenuation is usually frequency dependent (i.e., dispersion) causing the wave to change shape and spectral characteristics. So, Q methods may not be applicable because attenuation and dispersion are not included in the simulations used in this competition.
\nIt may be possible to adapt their general NN+Physics scheme/architecture as a way to include physics that is appropriate to this GWI situation, but that would take a bigger brain than mine 😄
", "votes": null, "replies": [ { "id": 3217377, "author_name": "billfan88", "author_url": "", "post_date": "06/05/2025 00:55:11", "content": "yeah. I always try to build a neural networks by the physics-based.
", "votes": null, "replies": [] }, { "id": 3217388, "author_name": "billfan88", "author_url": "", "post_date": "06/05/2025 01:17:02", "content": "