Page Not Found
Page not found. Your pixels are in another canvas.
Sitemap
A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.
Pages
Welcome
About
Posts
Blog Post number 1
Published:
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
publications
talks
An Open-source Object Storage for Distributed Acoustic Sensing
Published:
Distributed Acoustic Sensing (DAS) is the recording of strain measurements distributed along an optic fiber cable using photonic sensing. DAS can record signals from cars, trains, ships, planes, earthquakes, volcanic tremors, avalanches, footsteps, and whales. The technology is a revolution for seismology research, with great potential for frontier applications such as wildlife monitoring and the built environment. The data comes in at a high rate (100-10kHz) and multi-channels (100s-10,000s) continuously. The Photonic Sensing Facility (PSF) at UW already has 100sTBs of DAS data from 3 PSF-related experiments. At the same time, the PSF is planning several short- and long-term experiments that will bring the data archive to 1PB by the end of 2023. This large amount of data necessitates the exploration of the data format and metadata structure for fast data queries and processing. This incubator project aims to pilot the first cloud object storage to host DAS data. Our team will deploy a local cloud storage service on our local servers to emulate a cloud platform for DAS research using cloud-optimized data formats. The outcome of the project will be a pilot experiment promoted for authoritative seismic networks and archives.
Ensemble Learning for Seismic Phase Picking
Published:
The detection and picking of seismic waves are the first step toward earthquake catalog building, earthquake monitoring, and seismic hazard management. Recent advances in deep learning (DL) have leveraged the amount of labeled seismic data to improve the capability of detecting and picking earthquake signals. While these DL methods have shown great promise, their success remains hindered by low generalizability and poor performance in low signal-to-noise ratios (SNRs) data. Here, we propose a new processing workflow that integrates pre-trained DL models, multi-frequency band predictions, and ensemble estimations to enhance the generalization of these algorithms. We test the performance of the ensemble model using three benchmark datasets, one of which is within-domain and has been used for training the DL models, and the other two being cross-domain test datasets. We explore the performance given data and model characteristics. We also compare an ensemble approach with a transfer-learning approach and discuss the benefits and drawbacks of these two approaches when deploying on continuous data. Our experiments demonstrate that ensemble learning can drastically improve generalization ability and hence alleviate the need for transfer learning in the case where no labeled datasets exist.