<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>Automatic Media analysis: From monitoring to Insights</article-title>
      </title-group>
      <abstract>
        <p>One of the aspects of media analysis is brand monitoring which has traditionally focused on detecting and distributing every mention of a given brand in the media as quickly as possible. This monitoring aspect, which some companies such as Signal have automated using di erent IR and ML techniques, is critical for nding relevant information. However, retrieving these documents is not enough. Leaders in diverse organisations will use the knowledge derived from these relevant documents to change their decision making by understanding their risks, opportunities (e.g. Acquiring companies in distress or expanding to new territories), but this is not possible manually. Techniques to understand, organise and uncover the underlying insights that are hidden in the thousands of documents are needed. This talk will focus on the media analysis space, the current monitoring approach (using Signal as an example) and the future challenges and opportunities in the insights space, showcasing di erent challenges in the space and that the IR community could support with.</p>
      </abstract>
    </article-meta>
  </front>
  <body />
  <back>
    <ref-list />
  </back>
</article>