The Future Directions in Machine Learning workshop will be held 17-18 May 2017 at AFOSR/ONR in Arlington, VA. The goal of this workshop is identify critical technology areas for future investment to propel the state-of-the-art in machine learning (ML) toward a more general capability. To that end a panel of leading experts in machine learning and related fields has been assembled to discuss the state of the art in ML and provide insight on the underserved research topics.
The current state of practice in ML is heavily focused on supervised statistical methods requiring lots of labeled training data. While many successes have been achieved using this approach, it is not clear if this is sufficient to meet the full learning needs of future autonomous systems, or semi-autonomous systems working in human-machine teams. That is, our premise is that current ML is inadequate to support (semi-)autonomous systems operating continuously and persistently in highly dynamic domains requiring an ability to adapt to changes in environments for which they have not been “trained”, with little to no human interaction.
Because of time and space limitations, this workshop is by invitation only. If you did not receive an invitation directly, please contact the organizers at the email addresses below before attempting to register. Also, please note that open discussion – when observers will be allowed to ask questions to the panel – will be limited to 1 hour at the end of each day, to provide the distinguished panelists ample time to debate the hard problems.
Organizers:James Lawton (james.lawton.1@us.af.mil)Ben Knott (benjamin.knott.2@us.af.mil)Doug Riecken (richard.riecken@us.af.mil)
RegistrationClick here to register online.
AgendaClick here to view/download the agenda.
LocationThis workshop will be held in the Bobby Junker Executive Conference Center (BJECC) of the One Liberty Center Building located at 875 North Randolph Street., 14th Floor, Arlington, VA 22203.
Upon arrival, visitors must check in at the registration desk in the lobby. US federal government employees and/or military personnel displaying a valid CAC card or DoD building pass will be admitted as UNESCORTED guests. All other attendees, including US contractors will require an ESCORT. Photo identification will be verified and exchanged for a building badge upon check-in:
Parking* Limited parking is available in the building for $15.00 per day (Cash, Discover, Master Card and Visa) * Additional parking is available at the Ballston Common Mall
Local Area DiningClick here for information and map.