Mining a petabyte of seismic data | Cloud Conversation скачать в хорошем качестве

Mining a petabyte of seismic data | Cloud Conversation

Thanks to Yiyu Ni of the University of Washington for talking with us about a huge machine learning project building a new global earthquake catalog from over a petabyte of seismic data. Preprint paper on this work: https://arxiv.org/abs/2505.18874 More about our cloud efforts: https://www.earthscope.org/data/cloud/ 0:00 Scientific motivation 2:56 Data processing workflow 6:30 Adapting work to the cloud 9:28 Optimizing parallelization 13:50 Learning from the pilot test 17:05 Runtime and cost 22:10 Levers on cost 23:49 Picking the right number of CPUs 26:45 Learning from the main run 28:55 Applying learning to other projects 31:45 Reproducibility and collaboration 33:30 Advice for cloud veterans 36:00 Advice for cloud novices 37:28 Next steps for the project