Difference between revisions of "Machine Learning"

From Federation of Earth Science Information Partners
 
Line 1: Line 1:
 
__NOTOC__
 
__NOTOC__
 +
 
Machine learning, everybody’s doing it.  Scientists are applying machine learning to their scientific algorithms and using the results to justify various conclusions.  Is machine learning the silver bullet that will help us answer our scientific questions?
 
Machine learning, everybody’s doing it.  Scientists are applying machine learning to their scientific algorithms and using the results to justify various conclusions.  Is machine learning the silver bullet that will help us answer our scientific questions?
  
 
The purpose of this cluster is to educate ourselves and the ESIP community about machine learning through asking and answering questions and sharing experiences and resources.  The scope of the cluster includes topics like the following:
 
The purpose of this cluster is to educate ourselves and the ESIP community about machine learning through asking and answering questions and sharing experiences and resources.  The scope of the cluster includes topics like the following:
  
* What is machine learning?  What can machine learning do?  How is machine learning different from data science?  From data analytics?
+
*What is machine learning?  What can machine learning do?  How is machine learning different from data science?  From data analytics?
  
* What types of machine learning algorithms are there and how do they compare to each other?  Under what circumstances is each best applied and/or not applicable?
+
*What types of machine learning algorithms are there and how do they compare to each other?  Under what circumstances is each best applied and/or not applicable?
  
* What are symbolic and subsymbolic approaches to machine learning and their subtypes?
+
*What are symbolic and subsymbolic approaches to machine learning and their subtypes?
  
* What are machine learning SWOTs: strengths, weaknesses, opportunities, threats?    Some considerations include:
+
*What are machine learning SWOTs: strengths, weaknesses, opportunities, threats?    Some considerations include:
** models, model choices and biases
+
**models, model choices and biases
** p-hacking, data dredging
+
**p-hacking, data dredging
  
* What machine learning tools are available?
+
*What machine learning tools are available?
  
* In particular, regarding deep learning/neural networks
+
*In particular, regarding deep learning/neural networks
** How do supervised, semi-supervised, and unsupervised networks differ?
+
**How do supervised, semi-supervised, and unsupervised networks differ?
** How to integrate truth labeled data?
+
**How to integrate truth labeled data?
** Under what circumstances is it okay to not understand what the network learned?  When is it not okay?
+
**Under what circumstances is it okay to not understand what the network learned?  When is it not okay?
** How to deal with a lack of training data?
+
**How to deal with a lack of training data?
** How to (try to) understand what was learned?
+
**How to (try to) understand what was learned?
  
 
Possible cluster outputs could include:
 
Possible cluster outputs could include:
* A machine learning tool survey and SWOT analysis
 
* Training material, on line [and maybe a guide to available platforms and resources? e.g., using AWS ML platform]
 
* Training and/or other sessions at ESIP
 
* Recommendations regarding the application of ML in earth and space sciences
 
  
 +
*A machine learning tool survey and SWOT analysis
 +
*Training material, on line [and maybe a guide to available platforms and resources? e.g., using AWS ML platform]
 +
*Training and/or other sessions at ESIP
 +
*Recommendations regarding the application of ML in earth and space sciences
  
{| width="100%" cellpadding="0" cellspacing="0" style="zborder-top:1px solid #aaaaaa; border-collapse: collapse;"  
+
 
 +
{| style="zborder-top:1px solid #aaaaaa; border-collapse: collapse;" width="100%" cellspacing="0" cellpadding="0"  
 
|- valign="top" bgcolor="#FFFFFF"
 
|- valign="top" bgcolor="#FFFFFF"
|bgcolor="lightgreen" style="border: 1px solid gray;padding-left:0.5em;padding-right:0.5em;" width="50%"|
+
| style="border: 1px solid gray;padding-left:0.5em;padding-right:0.5em;" width="50%" bgcolor="lightgreen" |
 
===[[/Archived {{PAGENAME}} Events|News]]===
 
===[[/Archived {{PAGENAME}} Events|News]]===
  
Line 38: Line 40:
 
[[/Archived {{PAGENAME}} Events|Archive]]
 
[[/Archived {{PAGENAME}} Events|Archive]]
  
|bgcolor="lightblue" style="border: 1px solid gray;padding-left:0.5em;padding-right:0.5em;" width="50%"|
+
| style="border: 1px solid gray;padding-left:0.5em;padding-right:0.5em;" width="50%" bgcolor="lightblue" |
  
 
===Activities===
 
===Activities===
 
|}
 
|}
{| width="100%" cellpadding="0" cellspacing="0" style="zborder-top:1px solid #aaaaaa; border-collapse: collapse;"  
+
{| style="zborder-top:1px solid #aaaaaa; border-collapse: collapse;" width="100%" cellspacing="0" cellpadding="0"  
 +
 
 +
| style="border: 1px solid gray;padding-left:0.5em;padding-right:0.5em;" width="50%" bgcolor="#FFFFBB" |
  
|bgcolor="#FFFFBB" style="border: 1px solid gray;padding-left:0.5em;padding-right:0.5em;" width="50%"|
+
===Get Involved===
  
=== Get Involved===
+
*'''Email List''', https://lists.esipfed.org/mailman/listinfo/esip-machinelearning
* '''Email List''', [https://lists.esipfed.org/mailman/listinfo/esip-machinelearning https://lists.esipfed.org/mailman/listinfo/esip-machinelearning]
+
*'''Upcoming meetings'''
* '''Upcoming meetings'''
+
**Telecon: 3rd Friday of the month, 9:00PT/10:00MT/11:00CT/12:00ET, Access Code: 422-305-101
** Telecon: 3rd Friday of the month, 9:00PT/10:00MT/11:00CT/12:00ET, Access Code: 422-305-101
+
***[https://global.gotomeeting.com/join/422305101 GoToMeeting]
*** [https://global.gotomeeting.com/join/422305101     GoToMeeting]
+
***Dial In: United States: +1 (571) 317-3122
*** Dial In: United States: +1 (571) 317-3122  
 
  
* '''Contacts'''  
+
*'''Contacts'''  
 
**Anne Wilson, Founding Chair, Ronin Institute
 
**Anne Wilson, Founding Chair, Ronin Institute
 
**Ziheng Sun, Chair, George Mason University
 
**Ziheng Sun, Chair, George Mason University
Line 59: Line 62:
 
**Beth Huffer, Lingua Logica
 
**Beth Huffer, Lingua Logica
 
**Shawn Polson, Laboratory for Atmospheric and Space Physics (LASP)
 
**Shawn Polson, Laboratory for Atmospheric and Space Physics (LASP)
** Bill Teng, Goddard
+
**Bill Teng, Goddard
 
**Arif Albayrak, Goddard
 
**Arif Albayrak, Goddard
 
**Hook Hua, JPL
 
**Hook Hua, JPL
  
|bgcolor="pink" style="border: 1px solid gray;padding-left:0.5em;padding-right:0.5em;" width="50%"|
+
| style="border: 1px solid gray;padding-left:0.5em;padding-right:0.5em;" width="50%" bgcolor="pink" |
  
 
===Resources===
 
===Resources===
==== Current news ====
+
====Current news====
September 16, 2020, this cluster is initiating a [https://github.com/ESIPFed/Awesome-Earth-Artificial-Intelligence Awesome-Earth-Artificial-Intelligence] repository on Github. Calling for community contributions.
+
 
September 26, 2019, [https://www.nytimes.com/2019/09/26/technology/ai-computer-expense.html At Tech's Leading Edge, Worry About a Concentration of Power]
+
* September 16, 2020, this cluster is initiating a [https://github.com/ESIPFed/Awesome-Earth-Artificial-Intelligence Awesome-Earth-Artificial-Intelligence] repository on Github. Calling for community contributions.
:"The huge computing resources these companies have pose a threat — the universities cannot compete...”  "Academics are also raising concerns about the power consumed by advanced A.I. software. Training a large, deep-learning model can generate the same carbon footprint as the lifetime of five American cars, including gas, ..."
 
  
* July 2019, [https://arxiv.org/pdf/1907.10597.pdf Green AI]
+
* September 26, 2019, [https://www.nytimes.com/2019/09/26/technology/ai-computer-expense.html At Tech's Leading Edge, Worry About a Concentration of Power] "The huge computing resources these companies have pose a threat — the universities cannot compete...”  "Academics are also raising concerns about the power consumed by advanced A.I. software. Training a large, deep-learning model can generate the same carbon footprint as the lifetime of five American cars, including gas, ..."
  
*16 February 2019, [https://www.bbc.com/news/science-environment-47267081  AAAS: Machine learning 'causing science crisis']
+
*July 2019, [https://arxiv.org/pdf/1907.10597.pdf Green AI]
  
* November 7, 2018, [https://www.washingtonpost.com/opinions/chinas-application-of-ai-should-be-a-sputnik-moment-for-the-us-but-will-it-be/2018/11/06/69132de4-e204-11e8-b759-3d88a5ce9e19_story.html China's application of AI should be a Sputnik moment for the US.  But will it be?]
+
*16 February 2019, [https://www.bbc.com/news/science-environment-47267081 AAAS: Machine learning 'causing science crisis']
  
* August 21, 2018, [https://www.executivegov.com/2018/08/fy-2019-ndaa-to-authorize-10m-for-ai-national-security-commission/ FY 2019 NDAA to Authorize $10M for an AI National Security Commission]
+
*November 7, 2018, [https://www.washingtonpost.com/opinions/chinas-application-of-ai-should-be-a-sputnik-moment-for-the-us-but-will-it-be/2018/11/06/69132de4-e204-11e8-b759-3d88a5ce9e19_story.html China's application of AI should be a Sputnik moment for the US.  But will it be?]
 +
 
 +
*August 21, 2018, [https://www.executivegov.com/2018/08/fy-2019-ndaa-to-authorize-10m-for-ai-national-security-commission/ FY 2019 NDAA to Authorize $10M for an AI National Security Commission]
  
 
====Papers====
 
====Papers====
Line 85: Line 89:
  
 
|}
 
|}
 
  
 
[[category:CollabArea]]
 
[[category:CollabArea]]

Latest revision as of 21:19, September 16, 2020


Machine learning, everybody’s doing it. Scientists are applying machine learning to their scientific algorithms and using the results to justify various conclusions. Is machine learning the silver bullet that will help us answer our scientific questions?

The purpose of this cluster is to educate ourselves and the ESIP community about machine learning through asking and answering questions and sharing experiences and resources. The scope of the cluster includes topics like the following:

  • What is machine learning? What can machine learning do? How is machine learning different from data science? From data analytics?
  • What types of machine learning algorithms are there and how do they compare to each other? Under what circumstances is each best applied and/or not applicable?
  • What are symbolic and subsymbolic approaches to machine learning and their subtypes?
  • What are machine learning SWOTs: strengths, weaknesses, opportunities, threats? Some considerations include:
    • models, model choices and biases
    • p-hacking, data dredging
  • What machine learning tools are available?
  • In particular, regarding deep learning/neural networks
    • How do supervised, semi-supervised, and unsupervised networks differ?
    • How to integrate truth labeled data?
    • Under what circumstances is it okay to not understand what the network learned? When is it not okay?
    • How to deal with a lack of training data?
    • How to (try to) understand what was learned?

Possible cluster outputs could include:

  • A machine learning tool survey and SWOT analysis
  • Training material, on line [and maybe a guide to available platforms and resources? e.g., using AWS ML platform]
  • Training and/or other sessions at ESIP
  • Recommendations regarding the application of ML in earth and space sciences


News


Archive

Activities

Get Involved

  • Contacts
    • Anne Wilson, Founding Chair, Ronin Institute
    • Ziheng Sun, Chair, George Mason University
    • Yuhan Rao, Fellow, North Carolina Institute for Climate Studies (NCICS)
    • Beth Huffer, Lingua Logica
    • Shawn Polson, Laboratory for Atmospheric and Space Physics (LASP)
    • Bill Teng, Goddard
    • Arif Albayrak, Goddard
    • Hook Hua, JPL

Resources

Current news

  • September 26, 2019, At Tech's Leading Edge, Worry About a Concentration of Power "The huge computing resources these companies have pose a threat — the universities cannot compete...” "Academics are also raising concerns about the power consumed by advanced A.I. software. Training a large, deep-learning model can generate the same carbon footprint as the lifetime of five American cars, including gas, ..."

Papers