Working paper Open Access

Identifying AI talents among LinkedIn members, A machine learning approach

Thomas Roca


MARC21 XML Export

<?xml version='1.0' encoding='UTF-8'?>
<record xmlns="http://www.loc.gov/MARC21/slim">
  <leader>00000nam##2200000uu#4500</leader>
  <datafield tag="041" ind1=" " ind2=" ">
    <subfield code="a">eng</subfield>
  </datafield>
  <datafield tag="653" ind1=" " ind2=" ">
    <subfield code="a">Artificial Intelligence</subfield>
  </datafield>
  <datafield tag="653" ind1=" " ind2=" ">
    <subfield code="a">Skills</subfield>
  </datafield>
  <datafield tag="653" ind1=" " ind2=" ">
    <subfield code="a">Machine learning</subfield>
  </datafield>
  <datafield tag="653" ind1=" " ind2=" ">
    <subfield code="a">Natural Language Processing</subfield>
  </datafield>
  <datafield tag="653" ind1=" " ind2=" ">
    <subfield code="a">Big Data</subfield>
  </datafield>
  <controlfield tag="005">20190607190540.0</controlfield>
  <controlfield tag="001">3240963</controlfield>
  <datafield tag="856" ind1="4" ind2=" ">
    <subfield code="s">3583975</subfield>
    <subfield code="z">md5:cb98d32029907209e38510fa9f32990d</subfield>
    <subfield code="u">https://zenodo.org/record/3240963/files/AI in the Labour Force  Data for Policy June 2019.pdf</subfield>
  </datafield>
  <datafield tag="542" ind1=" " ind2=" ">
    <subfield code="l">open</subfield>
  </datafield>
  <datafield tag="260" ind1=" " ind2=" ">
    <subfield code="c">2019-04-23</subfield>
  </datafield>
  <datafield tag="909" ind1="C" ind2="O">
    <subfield code="p">openaire</subfield>
    <subfield code="p">user-dfp17</subfield>
    <subfield code="o">oai:zenodo.org:3240963</subfield>
  </datafield>
  <datafield tag="100" ind1=" " ind2=" ">
    <subfield code="u">Microsoft, Linkedin</subfield>
    <subfield code="a">Thomas Roca</subfield>
  </datafield>
  <datafield tag="245" ind1=" " ind2=" ">
    <subfield code="a">Identifying AI talents among LinkedIn members, A machine learning approach</subfield>
  </datafield>
  <datafield tag="980" ind1=" " ind2=" ">
    <subfield code="a">user-dfp17</subfield>
  </datafield>
  <datafield tag="540" ind1=" " ind2=" ">
    <subfield code="u">http://creativecommons.org/licenses/by/4.0/legalcode</subfield>
    <subfield code="a">Creative Commons Attribution 4.0 International</subfield>
  </datafield>
  <datafield tag="650" ind1="1" ind2="7">
    <subfield code="a">cc-by</subfield>
    <subfield code="2">opendefinition.org</subfield>
  </datafield>
  <datafield tag="520" ind1=" " ind2=" ">
    <subfield code="a">&lt;p&gt;How to identify specific profiles among the&amp;nbsp;hundred of millions gathered in LinkedIn?&amp;nbsp;LinkedIn Economic Graph thrives on skills,&lt;br&gt;
around 50 thousand of them are listed by&amp;nbsp;LinkedIn and constitute one of the main signals&amp;nbsp;to identify professions or trends. Artificial&amp;nbsp;Intelligence (AI) skills, for example, can be&amp;nbsp;used to identify the diffusion of AI in industries&amp;nbsp;[16]. But the noise can be loud around&amp;nbsp;skills for which the demand is high. Some&amp;nbsp;users may add &amp;quot;trendy&amp;quot; skills on their profiles&amp;nbsp;without having work experience or training&amp;nbsp;related to them. On the other hand, some&amp;nbsp;people may work in the broad AI ecosystem&amp;nbsp;(e.g. AI recruiters, AI sales&amp;nbsp;representatives,&amp;nbsp;etc.), without being the AI practitioners we&amp;nbsp;are looking for. Searching for keywords in profiles&amp;#39;&amp;nbsp;sections can lead to mis-identification of&amp;nbsp;certain profiles, especially for those related to&amp;nbsp;a field rather than an occupation. This is the&lt;br&gt;
case for Artificial Intelligence.&amp;nbsp;In this paper, we propose a machine learning&amp;nbsp;approach to identify such profiles, and suggest&lt;br&gt;
to train a binary text-classifier using job offers&amp;nbsp;posted on the platform rather than actual profiles.&lt;br&gt;
We suggest this approach allows to avoid&amp;nbsp;manually labeling the training dataset, granted&amp;nbsp;the assumption that job profiles posted by recruiters&amp;nbsp;are more &amp;quot;ideal-typical&amp;quot; or simply provide&amp;nbsp;a more consistent triptych &amp;quot;job title, job&amp;nbsp;description, associated skills&amp;quot; than the ones&amp;nbsp;that can be found among member&amp;#39;s profiles.&lt;/p&gt;</subfield>
  </datafield>
  <datafield tag="773" ind1=" " ind2=" ">
    <subfield code="n">doi</subfield>
    <subfield code="i">isVersionOf</subfield>
    <subfield code="a">10.5281/zenodo.2649207</subfield>
  </datafield>
  <datafield tag="024" ind1=" " ind2=" ">
    <subfield code="a">10.5281/zenodo.3240963</subfield>
    <subfield code="2">doi</subfield>
  </datafield>
  <datafield tag="980" ind1=" " ind2=" ">
    <subfield code="a">publication</subfield>
    <subfield code="b">workingpaper</subfield>
  </datafield>
</record>
190
128
views
downloads
All versions This version
Views 190134
Downloads 12875
Data volume 455.3 MB268.8 MB
Unique views 170124
Unique downloads 11769

Share

Cite as