<div dir="ltr"><div>The work seems to be a combination of AMDPs and the DAQN work that Mel and Chris are working on. An end to end task is broken into subtasks, each with an independent reward function and state-action space. They are learning each subtask independently and not all at the same time like DAQN. They chose a domain that Deepmind did for stacking blocks and showed that a policy for stack blocks can be learned 45x faster. Their video page does not exist so I can't really see their robot in action.</div><div>The paper itself lacks a lot of related work, they only cite the early options work, ignore everything after that and criticise two papers from last year. They do not point out that the hierarchy looks like a MAXQ hierarchy and do not talk about optimality issues of such an approach.</div><div>Best</div><div>nakul </div><div><br></div><div class="gmail_extra"><br><div class="gmail_quote">On 26 September 2017 at 15:00, Littman, Michael <span dir="ltr"><<a href="mailto:mlittman@cs.brown.edu" target="_blank">mlittman@cs.brown.edu</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">indeed... has anyone looked over the paper? any insights?</div><div class="gmail_extra"><br><div class="gmail_quote"><div><div class="m_1297300274192081933h5">On Tue, Sep 26, 2017 at 2:48 PM, Marie desJardins <span dir="ltr"><<a href="mailto:mariedj@umbc.edu" target="_blank">mariedj@umbc.edu</a>></span> wrote:<br></div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div class="m_1297300274192081933h5">
<div text="#000000" bgcolor="#FFFFFF">
<font face="Lucida Grande">Looks an awful lot like AMDPs...<br>
<br>
<a class="m_1297300274192081933m_2755631248327476707m_-7242337529551291957moz-txt-link-freetext" href="https://www.forbes.com/sites/aarontilley/2017/09/19/ai-startup-invents-trick-for-robots-to-more-efficiently-teach-themselves-complex-tasks/#5a6254d315fe" target="_blank">https://www.forbes.com/sites/a<wbr>arontilley/2017/09/19/ai-start<wbr>up-invents-trick-for-robots-to<wbr>-more-efficiently-teach-themse<wbr>lves-complex-tasks/#5a6254d315<wbr>fe</a><br>
<br>
<a class="m_1297300274192081933m_2755631248327476707m_-7242337529551291957moz-txt-link-freetext" href="https://drive.google.com/file/d/0B7-VbSZ5FzXBdURHLXV4OU9EOTQ/view" target="_blank">https://drive.google.com/file/<wbr>d/0B7-VbSZ5FzXBdURHLXV4OU9EOTQ<wbr>/view</a><span class="m_1297300274192081933m_2755631248327476707HOEnZb"><font color="#888888"><br>
<br>
Marie<br>
<br>
</font></span></font><span class="m_1297300274192081933m_2755631248327476707HOEnZb"><font color="#888888">
<div class="m_1297300274192081933m_2755631248327476707m_-7242337529551291957moz-signature">-- <br>
Dr. Marie desJardins
<br>
Associate Dean for Academic Affairs
<br>
College of Engineering and Information Technology
<br>
University of Maryland, Baltimore County
<br>
1000 Hilltop Circle
<br>
Baltimore MD 21250
<br>
<br>
Email: <a class="m_1297300274192081933m_2755631248327476707m_-7242337529551291957moz-txt-link-abbreviated" href="mailto:mariedj@umbc.edu" target="_blank">mariedj@umbc.edu</a>
<br>
Voice: <a href="tel:(410)%20455-3967" value="+14104553967" target="_blank">410-455-3967</a>
<br>
Fax: <a href="tel:(410)%20455-3559" value="+14104553559" target="_blank">410-455-3559</a></div>
</font></span></div>
<br></div></div>______________________________<wbr>_________________<br>
Robot-learning mailing list<br>
<a href="mailto:Robot-learning@cs.umbc.edu" target="_blank">Robot-learning@cs.umbc.edu</a><br>
<a href="https://lists.cs.umbc.edu/mailman/listinfo/robot-learning" rel="noreferrer" target="_blank">https://lists.cs.umbc.edu/mail<wbr>man/listinfo/robot-learning</a><br>
<br></blockquote></div><br></div>
<br>______________________________<wbr>_________________<br>
Robot-learning mailing list<br>
<a href="mailto:Robot-learning@cs.umbc.edu" target="_blank">Robot-learning@cs.umbc.edu</a><br>
<a href="https://lists.cs.umbc.edu/mailman/listinfo/robot-learning" rel="noreferrer" target="_blank">https://lists.cs.umbc.edu/mail<wbr>man/listinfo/robot-learning</a><br>
<br></blockquote></div><br></div></div>