BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20191206T142452Z
LOCATION:Analog 1\, 2
DTSTART;TZID=Europe/Stockholm:20180625T132900
DTEND;TZID=Europe/Stockholm:20180625T133300
UID:isc_hpc_ISC High Performance 2018_sess103_phd115@linklings.com
SUMMARY:(PhD08) A Machine Learning Workflow for Hurricane Prediction
DESCRIPTION:PhD Forum\n\n(PhD08) A Machine Learning Workflow for Hurricane
  Prediction\n\nKahira, Bautista-Gomez, Badia\n\nThe Atlantic hurricane sea
 son runs from June 1st to November causing massive destruction and loss of
  life. In  2017, 17 named storms hit the Atlantic causing destruction wort
 h an estimated $316 million and at least 464 fatalities. Meteorologists, b
 y studying previous weather data, predict the expected number of hurricane
 s in the season. These predictions help authorities prepare for disasters 
 and over the years, better predictions have minimized loss of life and pro
 perty. However, these predictions rely on human expertise and are often ex
 tremely complex due to the thousands of parameters involved and the chaoti
 c nature of weather. \n\nThe aim of this study is to introduce Machine Lea
 rning to hurricane prediction. Recent scientific advancements have seen ge
 ostationary satellites capable of collecting tens of Terabytes of daily da
 ta of the weather. On the other side, machine learning models propose effi
 cient techniques to analyze such large data and extract meaningful informa
 tion. With this large amount of data and the power of high performance com
 puting, machine learning could be an alternative tool for climate study.\n
 \nMachine learning has the ability to understand complex models and relati
 onships in data. Recent developments in deep learning models such as Deep 
 Neural Networks (DNN) have led to significant achievements in accuracy. We
  introduce machine learning model to hurricane prediction to explore the c
 omplex relationship between multiple factors such as sea surface temperatu
 re, sea level pressure, sea ice cover and wind patterns. We aim to apply a
  deep learning model to understand the effect of these parameters in the h
 urricane season and the number of hurricanes. Such insights could signific
 antly improve disaster preparedness and give authorities a better picture 
 of what to expect in the hurricane season.\n\nUsing historical data, we tr
 ain a DNN to classify the hurricane season based on the number of hurrican
 es likely to occur. Our initial experiments are aimed at finding the most 
 significant factors in storm formation. Preliminary results show that Sea 
 surface temperature has the highest impact on the prediction of the number
  of storms. Furthermore, given the average sea surface temperatures in a m
 onth, our DNN model is able to predict the number of hurricanes with about
  60% of accuracy.\n\nOur future goal is to develop a complete end to end w
 orkflow to continuously learn weather patterns that affect the hurricane s
 eason and accordingly, to make predictions. We plan to implement distribut
 ed learning using PyCOMPSs (a programming model and runtime which aims to 
 ease the development of parallel applications for distributed infrastructu
 res, such as Clusters and Clouds) to reduce or eliminate the need to expen
 sive computational infrastructure in climate science.\n\nPasses: Conferenc
 e Pass, Exhibition Pass, AI/Machine Learning/Deep Learning, Big Data Analy
 tics\n\nTag: Conference Pass, Exhibition Pass, AI/Machine Learning/Deep Le
 arning, Big Data Analytics
URL:https://2018.isc-program.com/presentation/?id=phd115&sess=sess103
END:VEVENT
END:VCALENDAR