<div dir="ltr"><div class="gmail_quote"><div dir="ltr"><div class="gmail_quote"><div><font size="4"><b style="color:rgb(255,0,0)">"</b><font color="#ff0000"><span><b><div style="display:inline" class="gmail_default">Earthquakes and Artificial Intelligence<font color="#ff0000"><span><b>"</b></span></font></div></b></span></font></font></div><div><font size="2" color="#ff0000"><span><b><br></b></span></font></div><font size="2"><b><div style="display:inline" class="gmail_default">Zachary Ross<br></div></b><i><div style="display:inline" class="gmail_default">Caltech Seismological Laboratory<span style="color:rgb(0,0,0)"> <br></span></div></i></font></div><div class="gmail_quote"><div class="gmail_quote"><font size="2" color="#000000"><br></font></div><div class="gmail_quote">
<div class="gmail-page" title="Page 1">
<div class="gmail-layoutArea">
<div class="gmail-column">
<p><font size="2"><span>With the volume of seismic data recorded around the world growing rapidly by the year, reliable
automated techniques are becoming increasingly utilized for extracting the usable scientific signal. Such
automated methods, however, generally lag behind in performance compared with that of expert
seismologists. To address these challenges, we develop a framework for generalized seismic phase
detection using deep learning, employing algorithms from artificial intelligence that learn directly from
seismograms without the need for feature extraction. We utilize millions of manually determined phase
picks from the Southern California Seismic Network to train deep neural networks to detect P and S
body waves, localize onset times, and determine first-motion polarities for calculating focal
mechanisms. The networks are shown to perform as well or better than professional seismic analysts
due to their ability to reliably generalize entire waveform and meta-data archives. This enables the
trained models to be applied to data from completely different tectonic regimes or to large magnitude
events without representation in the training set. We demonstrate the detection capabilities of the
method on the 2016 Bombay Beach, California swarm and the 2016 Mw 7.0 Kumamoto earthquake, and
additionally demonstrate the picking capabilities by building a fully-automated focal mechanism
catalog for southern California. We separately apply deep learning to the challenging task of
earthquake phase association and demonstrate its potential for robust, grid-free phase association even
in the presence of numerous overlapping events. The developed techniques have important
applications to real-time seismology and Earthquake Early warning.
</span></font></p>
</div>
</div>
</div>
</div><font size="2"><br><b><div style="display:inline"><div style="display:inline">Thursday, <div style="display:inline">July <div style="display:inline" class="gmail_default">26</div>th</div></div></div></b><b>, 2018</b><br><b><span><div style="display:inline"><div style="display:inline"><div style="display:inline"><div style="display:inline"><div style="display:inline">10</div><div style="display:inline"></div><div style="display:inline">a</div>m<div style="display:inline"></div></div></div></div></div></span></b><b><span></span></b><b><span><div style="display:inline"><div style="display:inline"><div style="display:inline"></div> </div></div> (Mountain Time)</span></b><br><b><span>USGS, 1711 Illinois Street, Golden, CO</span></b><br><b><span>Entry Level <span class="gmail-m_4312014451518659657gmail-m_642874087160079801m_8169915678037529868gmail-m_-1429313160177607069gmail-m_2369097616388194495gmail-m_1479565434165998917m_7066959450357385387m_-8586038301793833330m_2174399009776966998gmail-m_3536024389969136014gmail-m_6564323528617604720gmail-m_-3202867471772006187m_-3474394390109531423m_-6180755618559776416m_-6692829955733685287gmail-m_-1920778670233885335gmail-m_2928417918433606341m_548290485840828435m_3628773897655229900gmail-m_7446487874202752399m_-8264889372741220741gmail-m_5431511284334585236gmail-m_-1882990626170291808m_9106775657748341738gmail-m_-1921791487083127966gmail-m_4140643564653640769m_-4556382912284060029m_2261340191774375389gmail-m_3025069362589959092gmail-m_-8968070735390492279m_4896026561608388806gmail-m_3918809983636524111m_7068817575649352729m_-6172873775242136948gmail-m_-4303968599703518853gmail-m_2043882230405221244gmail-m_2480559937188047531m_-1887565490609744462m_8131614935401267798m_6252585815979993878gmail-m_-5616756539997506461gmail-m_3957397034944864967gmail-m_-5590660146057455466gmail-m_2010942344340859242gmail-m_4771098111974711889gmail-m_337406015407620904gmail-m_-5330926199930091226m_647861723187015620gmail-il"><span class="gmail-m_4312014451518659657gmail-m_642874087160079801m_8169915678037529868gmail-m_-1429313160177607069gmail-m_2369097616388194495gmail-m_1479565434165998917m_7066959450357385387m_-8586038301793833330m_2174399009776966998gmail-m_3536024389969136014gmail-m_6564323528617604720gmail-m_-3202867471772006187m_-3474394390109531423m_-6180755618559776416m_-6692829955733685287gmail-m_-1920778670233885335gmail-m_2928417918433606341m_548290485840828435m_3628773897655229900gmail-m_7446487874202752399m_-8264889372741220741gmail-m_5431511284334585236gmail-m_-1882990626170291808m_9106775657748341738gmail-m_-1921791487083127966gmail-m_4140643564653640769m_-4556382912284060029m_2261340191774375389gmail-m_3025069362589959092gmail-m_-8968070735390492279m_4896026561608388806gmail-m_3918809983636524111m_7068817575649352729m_-6172873775242136948gmail-m_-4303968599703518853gmail-m_2043882230405221244gmail-m_2480559937188047531m_-1887565490609744462m_8131614935401267798m_6252585815979993878gmail-m_-5616756539997506461gmail-m_3957397034944864967gmail-m_-5590660146057455466gmail-m_2010942344340859242gmail-m_4771098111974711889gmail-il"><span class="gmail-m_4312014451518659657gmail-m_642874087160079801m_8169915678037529868gmail-m_-1429313160177607069gmail-il">Seminar</span></span></span> Room (204)</span></b><br><span><br></span><span>*Note: Please arrive <u>~5 minutes early</u> and <u>bring photo ID</u> for airport-style </span><span>security</span><span> measures now in place at the USGS building.</span><br><span><br></span>Thank you,<br><span class="gmail-m_4312014451518659657gmail-m_642874087160079801m_8169915678037529868gmail-m_-1429313160177607069gmail-m_2369097616388194495gmail-m_1479565434165998917m_7066959450357385387m_-8586038301793833330m_2174399009776966998gmail-m_3536024389969136014gmail-m_6564323528617604720gmail-m_-3202867471772006187m_-3474394390109531423m_-6180755618559776416m_-6692829955733685287gmail-m_-1920778670233885335gmail-m_2928417918433606341m_548290485840828435m_3628773897655229900gmail-m_7446487874202752399m_-8264889372741220741gmail-m_5431511284334585236gmail-m_-1882990626170291808m_9106775657748341738gmail-m_6925426022411937115m_7055967733504661449gmail-m_-8968070735390492279m_4896026561608388806gmail-m_3918809983636524111m_7068817575649352729m_-6172873775242136948gmail-m_-4303968599703518853gmail-m_2043882230405221244gmail-m_2480559937188047531m_-1887565490609744462m_8131614935401267798m_6252585815979993878gmail-m_-5616756539997506461gmail-m_3957397034944864967gmail-m_-5590660146057455466gmail-m_2010942344340859242gmail-m_4771098111974711889gmail-il">GHSC</span> <span class="gmail-m_4312014451518659657gmail-m_642874087160079801m_8169915678037529868gmail-m_-1429313160177607069gmail-m_2369097616388194495gmail-m_1479565434165998917m_7066959450357385387m_-8586038301793833330m_2174399009776966998gmail-m_3536024389969136014gmail-m_6564323528617604720gmail-m_-3202867471772006187m_-3474394390109531423m_-6180755618559776416m_-6692829955733685287gmail-m_-1920778670233885335gmail-m_2928417918433606341m_548290485840828435m_3628773897655229900gmail-m_7446487874202752399m_-8264889372741220741gmail-m_5431511284334585236gmail-m_-1882990626170291808m_9106775657748341738gmail-m_6925426022411937115m_7055967733504661449gmail-m_-8968070735390492279m_4896026561608388806gmail-m_3918809983636524111m_7068817575649352729m_-6172873775242136948gmail-m_-4303968599703518853gmail-m_2043882230405221244gmail-m_2480559937188047531m_-1887565490609744462m_8131614935401267798m_6252585815979993878gmail-m_-5616756539997506461gmail-m_3957397034944864967gmail-m_-5590660146057455466gmail-m_2010942344340859242gmail-m_4771098111974711889gmail-m_337406015407620904gmail-m_-5330926199930091226m_647861723187015620gmail-il">Seminar</span> Committee <br><span><br></span><span>Mirus, Ben - </span><a href="mailto:bbmirus@usgs.gov" target="_blank">bbmirus@usgs.gov</a><br>Josh Rigler - <a href="mailto:erigler@usgs.gov" target="_blank">erigler@usgs.gov</a><br></font><div><div class="gmail_quote"><div><font size="2">Oliver Boyd - <a href="mailto:oboyd@usgs.gov" target="_blank">olboyd@usgs.gov</a></font></div></div></div>
</div></div>
</div></div>