{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,16]],"date-time":"2026-10-16T04:43:47Z","timestamp":1792125827907,"version":"4.4.0"},"reference-count":47,"publisher":"Oxford University Press (OUP)","issue":"10","license":[{"start":{"date-parts":[[2026,9,18]],"date-time":"2026-09-18T00:00:00Z","timestamp":1789689600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"funder":[{"name":"Data Institute for Societal Challenges"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,9,30]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Seismic monitoring is entering a high data volume, multimodal sensing regime driven by dense sensor deployments, including micro-electro-mechanical systems arrays, large-N (100s\u20131000s) Nodal arrays, distributed acoustic sensing, and other geophysical instruments. Machine learning models, with appropriate tuning and calibration, are capable of achieving close to expert-level performance in seismic detection and phase picking. The emerging operational challenge is how to scale that machine intelligence across future multimodal sensing networks. Effective incorporation of these capabilities remains constrained by telemetry bandwidth, centralized computation architectures, and communication latency. Here, we present XiaoNet, a proof-of-concept edge machine learning architecture for seismicity monitoring based on a simple system principle: perform first-pass inference near the sensor and transmit concise metadata rather than continuous streams. Using knowledge distillation, we compress a state-of-the-art PhaseNet model by over 93% into XiaoNet, a 240\u2005KB model designed for edge deployment, while preserving strong phase-picking performance across datasets. Deployment on a Raspberry Pi 5 shows sub-millisecond inference latency for 30-s 3-channel waveform windows. By shifting inference to the edge and transmitting concise event metadata payloads, the framework demonstrated the potential for edge inference to reduce communication latency. Substituting an event metadata message payload for continuous raw waveform stream largely mitigates the dominant source of variable telemetry delay under moderate to heavy load. This result supports a feasible architecture in which edge inference devices complements centralized observatories, potentially improving resilience and response speed for seismic monitoring and may extend to future multimodal subsurface and infrastructure sensing systems.<\/jats:p>","DOI":"10.1093\/pnasnexus\/pgag322","type":"journal-article","created":{"date-parts":[[2026,9,18]],"date-time":"2026-09-18T14:54:36Z","timestamp":1789743276000},"source":"Crossref","is-referenced-by-count":0,"title":["Edge seismic phase picking: Model distillation for multimodal early-warning systems"],"prefix":"10.1093","volume":"5","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-5403-4650","authenticated-orcid":false,"given":"Hongyu","family":"Xiao","sequence":"first","affiliation":[{"name":"Oklahoma Geological Survey, University of Oklahoma , Norman, OK 73019 ,","place":["USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7127-9422","authenticated-orcid":false,"given":"Jacob I","family":"Walter","sequence":"additional","affiliation":[{"name":"Oklahoma Geological Survey, University of Oklahoma , Norman, OK 73019 ,","place":["USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6234-6910","authenticated-orcid":false,"given":"Paul","family":"Ogwari","sequence":"additional","affiliation":[{"name":"Oklahoma Geological Survey, University of Oklahoma , Norman, OK 73019 ,","place":["USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2026,9,18]]},"reference":[{"key":"2026101523463591700_pgag322-B1","doi-asserted-by":"crossref","first-page":"2894","DOI":"10.1785\/0120180080","article-title":"Generalized seismic phase detection with deep learning","volume":"108","author":"Ross","year":"2018","journal-title":"Bull Seismol Soc Am"},{"key":"2026101523463591700_pgag322-B2","first-page":"261","article-title":"PhaseNet: a deep-neural-network-based seismic arrival-time picking method","volume":"216","author":"Zhu","year":"2019","journal-title":"Geophys J Int"},{"key":"2026101523463591700_pgag322-B3","doi-asserted-by":"crossref","first-page":"3952","DOI":"10.1038\/s41467-020-17591-w","article-title":"Earthquake transformer\u2014an attentive deep-learning model for simultaneous earthquake detection and phase picking","volume":"11","author":"Mousavi","year":"2020","journal-title":"Nat Commun"},{"key":"2026101523463591700_pgag322-B4","doi-asserted-by":"crossref","DOI":"10.1029\/2020JB021473","article-title":"Scalodeep: a highly generalized deep learning framework for real-time earthquake detection","volume":"126","author":"Saad","year":"2021","journal-title":"J Geophys Res Solid Earth"},{"key":"2026101523463591700_pgag322-B5","first-page":"555","article-title":"easyQuake: putting machine learning to work for your regional seismic network or local earthquake study","volume":"92","author":"Walter","year":"2021","journal-title":"Seismol Soc Am"},{"key":"2026101523463591700_pgag322-B6","article-title":"Onsite early prediction of PGA using CNN with multi-scale and multi-domain P-waves as input","volume":"9","author":"Hsu","year":"2021","journal-title":"Front Earth Sci (Lausanne)"},{"key":"2026101523463591700_pgag322-B7","doi-asserted-by":"crossref","DOI":"10.1029\/2021JB023842","article-title":"3D microseismic monitoring using machine learning","volume":"127","author":"Chen","year":"2022","journal-title":"J Geophys Res Solid Earth"},{"key":"2026101523463591700_pgag322-B8","doi-asserted-by":"crossref","first-page":"611","DOI":"10.1785\/0220190211","article-title":"The Oklahoma geological survey statewide seismic network","volume":"91","author":"Walter","year":"2020","journal-title":"Seismol Res Lett"},{"key":"2026101523463591700_pgag322-B9","doi-asserted-by":"crossref","first-page":"2013","DOI":"10.1785\/0220230327","article-title":"TXED: the Texas earthquake dataset for AI","volume":"95","author":"Chen","year":"2024","journal-title":"Seismol Res Lett"},{"key":"2026101523463591700_pgag322-B10","doi-asserted-by":"crossref","DOI":"10.1126\/science.1225942","article-title":"Injection-induced earthquakes","volume":"341","author":"Ellsworth","year":"2013","journal-title":"Science"},{"key":"2026101523463591700_pgag322-B11","doi-asserted-by":"crossref","first-page":"2437","DOI":"10.1785\/0220180105","article-title":"Natural and induced seismicity in the Texas and Oklahoma Panhandles","volume":"89","author":"Walter","year":"2018","journal-title":"Seismol Res Lett"},{"key":"2026101523463591700_pgag322-B12","doi-asserted-by":"crossref","first-page":"1007","DOI":"10.1126\/science.aat6081","article-title":"Assessing whether the 2017M w 5.4 Pohang earthquake in South Korea was an induced event","volume":"360","author":"Kim","year":"2018","journal-title":"Science"},{"key":"2026101523463591700_pgag322-B13","doi-asserted-by":"crossref","first-page":"2695","DOI":"10.1785\/0220250170","article-title":"Hydraulic fracturing enhances wastewater disposal seismicity in eastern Oklahoma","volume":"97","author":"Ogwari","year":"2025","journal-title":"Seismol Res Lett"},{"key":"2026101523463591700_pgag322-B14","doi-asserted-by":"publisher","volume-title":"Recent seismicity in the Southern Midcontinent, USA: scientific, regulatory, and industry responses: Geological Society of America Special Paper 559","author":"Murray","DOI":"10.1130\/2023.2559(02)"},{"key":"2026101523463591700_pgag322-B15","doi-asserted-by":"crossref","first-page":"26","DOI":"10.1785\/gssrl.80.1.26","article-title":"The quake-catcher network: citizen science expanding seismic horizons","volume":"80","author":"Cochran","year":"2009","journal-title":"Seismol Res Lett"},{"key":"2026101523463591700_pgag322-B16","doi-asserted-by":"crossref","DOI":"10.1126\/sciadv.1501055","article-title":"Myshake: a smartphone seismic network for earthquake early warning and beyond","volume":"2","author":"Kong","year":"2016","journal-title":"Sci Adv"},{"key":"2026101523463591700_pgag322-B17","doi-asserted-by":"crossref","first-page":"32","DOI":"10.1016\/j.ijdrr.2018.02.024","article-title":"LastQuake: from rapid information to global seismic risk reduction","volume":"28","author":"Bossu","year":"2018","journal-title":"Int J Disaster Risk Reduct"},{"key":"2026101523463591700_pgag322-B18","doi-asserted-by":"crossref","first-page":"3093","DOI":"10.3390\/s19143093","article-title":"A review of the capacitive MEMS for seismology","volume":"19","author":"D\u2019Alessandro","year":"2019","journal-title":"Sensors"},{"key":"2026101523463591700_pgag322-B19","doi-asserted-by":"crossref","first-page":"219","DOI":"10.1785\/0220180251","article-title":"Do low-cost seismographs perform well enough for your network? An overview of laboratory tests and field observations of the OSOP Raspberry Shake 4D","volume":"90","author":"Anthony","year":"2019","journal-title":"Seismol Res Lett"},{"key":"2026101523463591700_pgag322-B20","doi-asserted-by":"crossref","first-page":"2963","DOI":"10.3390\/s20102963","article-title":"BLESeis: low-cost IOT sensor for smart earthquake detection and notification","volume":"20","author":"Won","year":"2020","journal-title":"Sensors"},{"key":"2026101523463591700_pgag322-B21","doi-asserted-by":"crossref","first-page":"1328","DOI":"10.1038\/s41598-018-36675-8","article-title":"Distributed acoustic sensing using dark fiber for near-surface characterization and broadband seismic event detection","volume":"9","author":"Ajo-Franklin","year":"2019","journal-title":"Sci Rep"},{"key":"2026101523463591700_pgag322-B22","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1785\/0220190112","article-title":"Distributed acoustic sensing turns fiber-optic cables into sensitive seismic antennas","volume":"91","author":"Zhan","year":"2020","journal-title":"Seismol Res Lett"},{"key":"2026101523463591700_pgag322-B23","doi-asserted-by":"crossref","first-page":"946","DOI":"10.1785\/0220130043","article-title":"Optimal seismic network density for earthquake early warning: a case study from California","volume":"84","author":"Kuyuk","year":"2013","journal-title":"Seismol Res Lett"},{"key":"2026101523463591700_pgag322-B24","doi-asserted-by":"crossref","first-page":"361","DOI":"10.1146\/annurev-earth-053018-060457","article-title":"Earthquake early warning: advances, scientific challenges, and societal needs","volume":"47","author":"Allen","year":"2019","journal-title":"Annu Rev Earth Planet Sci"},{"key":"2026101523463591700_pgag322-B25","first-page":"170","article-title":"Latency of waveform data delivery from the southern California seismic network during the 2019 Ridgecrest earthquake sequence and its effect on ShakeAlert","volume":"92","author":"Stubailo","year":"2021","journal-title":"Seismol Soc Am"},{"key":"2026101523463591700_pgag322-B26","author":"Ba"},{"key":"2026101523463591700_pgag322-B27","doi-asserted-by":"crossref","first-page":"3289","DOI":"10.1785\/0220250354","article-title":"Evaluation of distributed acoustic sensing phase pick quality and performance for operational earthquake monitoring","volume":"97","author":"Tepp","year":"2026","journal-title":"Seismol Res Lett"},{"key":"2026101523463591700_pgag322-B28","doi-asserted-by":"crossref","first-page":"28938","DOI":"10.1038\/s41598-025-14491-1","article-title":"Improvements from incorporating machine learning algorithms into near real-time operational post-processing","volume":"15","author":"Tepp","year":"2025","journal-title":"Sci Rep"},{"key":"2026101523463591700_pgag322-B29","doi-asserted-by":"crossref","first-page":"960","DOI":"10.1038\/s41598-025-30568-3","article-title":"Integrating fiber-optic seismic arrays into earthquake early warning systems with the dEPIC framework","volume":"16","author":"Gou","year":"2025","journal-title":"Sci Rep"},{"key":"2026101523463591700_pgag322-B30","doi-asserted-by":"crossref","first-page":"524","DOI":"10.1785\/0120220182","article-title":"Earthquake phase association with graph neural networks","volume":"113","author":"McBrearty","year":"2023","journal-title":"Bull Seismol Soc Am"},{"key":"2026101523463591700_pgag322-B31","doi-asserted-by":"crossref","DOI":"10.1029\/2023GL106434","article-title":"Phase neural operator for multi-station picking of seismic arrivals","volume":"50","author":"Sun","year":"2023","journal-title":"Geophys Res Lett"},{"key":"2026101523463591700_pgag322-B32","author":"Zhao"},{"key":"2026101523463591700_pgag322-B33","first-page":"2030","article-title":"Earthquake detection with tinyML","volume":"94","author":"Clements","year":"2023","journal-title":"Seismol Soc Am"},{"key":"2026101523463591700_pgag322-B34","doi-asserted-by":"crossref","first-page":"105","DOI":"10.1146\/annurev-earth-071822-100323","article-title":"Machine learning in earthquake seismology","volume":"51","author":"Mousavi","year":"2023","journal-title":"Annu Rev Earth Planet Sci"},{"key":"2026101523463591700_pgag322-B35","author":"Bucilu\u01ce"},{"key":"2026101523463591700_pgag322-B36","author":"Hinton"},{"key":"2026101523463591700_pgag322-B37","doi-asserted-by":"publisher","author":"Howard","year":"17  2017","DOI":"10.48550\/arXiv.1704.04861"},{"key":"2026101523463591700_pgag322-B38","doi-asserted-by":"crossref","first-page":"179464","DOI":"10.1109\/ACCESS.2019.2947848","article-title":"STanford EArthquake Dataset (STEAD): a global data set of seismic signals for AI","volume":"7","author":"Mousavi","year":"2019","journal-title":"IEEE Access"},{"key":"2026101523463591700_pgag322-B39","volume-title":"National seismic networks of Switzerland","author":"Swiss Seismological Service (SED) At ETH Zurich","year":"1983"},{"key":"2026101523463591700_pgag322-B40","article-title":"Transfer learning and benchmarking for induced seismic event detection: insights from Oklahoma","volume":"3","author":"Xiao","year":"2026","journal-title":"J Geophys Res Mach Learn Comput"},{"key":"2026101523463591700_pgag322-B41","author":"Banks, A"},{"key":"2026101523463591700_pgag322-B42","first-page":"166","article-title":"The generalization ability research of deep learning algorithm in seismic phase detection of regional seismic network","volume":"41","author":"Zhao","year":"2021","journal-title":"Earthquake"},{"key":"2026101523463591700_pgag322-B43","doi-asserted-by":"crossref","first-page":"155","DOI":"10.1007\/s10462-025-11157-2","article-title":"Application of artificial intelligence technology in the study of anthropogenic earthquakes: a review","volume":"58","author":"Li","year":"2025","journal-title":"Artif Intell Rev"},{"key":"2026101523463591700_pgag322-B44","first-page":"1273","volume-title":"Artificial intelligence and statistics","author":"McMahan","year":"2017"},{"key":"2026101523463591700_pgag322-B45","doi-asserted-by":"crossref","first-page":"8192","DOI":"10.1038\/s41467-023-43355-3","article-title":"Seismic arrival-time picking on distributed acoustic sensing data using semi-supervised learning","volume":"14","author":"Zhu","year":"2023","journal-title":"Nat Commun"},{"key":"2026101523463591700_pgag322-B46","first-page":"1695","article-title":"SeisBench\u2014a toolbox for machine learning in seismology","volume":"93","author":"Woollam","year":"2022","journal-title":"Seismol Soc Am"},{"key":"2026101523463591700_pgag322-B47","doi-asserted-by":"publisher","author":"Howard","DOI":"10.1109\/ICCV.2019.00140"}],"container-title":["PNAS Nexus"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/pnasnexus\/advance-article-pdf\/doi\/10.1093\/pnasnexus\/pgag322\/71225997\/pgag322.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/pnasnexus\/article-pdf\/5\/10\/pgag322\/71225997\/pgag322.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/pnasnexus\/article-pdf\/5\/10\/pgag322\/71225997\/pgag322.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,16]],"date-time":"2026-10-16T03:46:49Z","timestamp":1792122409000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/pnasnexus\/article\/doi\/10.1093\/pnasnexus\/pgag322\/8817284"}},"subtitle":[],"editor":[{"given":"Barbara","family":"Romanowicz","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,9,18]]},"references-count":47,"journal-issue":{"issue":"10","published-print":{"date-parts":[[2026,9,30]]}},"URL":"https:\/\/doi.org\/10.1093\/pnasnexus\/pgag322","relation":{},"ISSN":["2752-6542"],"issn-type":[{"value":"2752-6542","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,10]]},"published":{"date-parts":[[2026,9,18]]},"article-number":"pgag322"}}