OVERVIEW

 A human-level artificially intelligent agent must be able to represent 
 and reason about the world, at some level, in terms of high-level 
 concepts such as entities and relations. The problem of acquiring these 
 rich high-level representations, known as the "knowledge acquisition 
 bottleneck", has long been an obstacle for achieving human-level AI. A 
 popular approach to this problem is to handcraft these high-level 
 representations, but this has had limited success. An alternate approach 
 is for rich representations to be learned autonomously from low-level 
 sensor data. Potentially, the latter approach may yield more robust 
 representations, and should rely less on human knowledge-engineering.


 TOPICS

 We are interested in all parts of the bridge between low-level-sensors 
 and rich high-level representations and their use in reasoning tasks.

 - Learning concept hierarchies from sensor data.
 - Representing and learning invariant concepts.
 - Postulating objects and theoretical entities.
 - Postulating relations from sensor data, when the data is not 
 explicitly relational.
 - Learning symbolic representations from numerical sensor data.
 - High-level reasoning grounded in robotic sensors and effectors.
 - Sensor-grounded research on cognitive architectures.

 Although we are most interested in general learning methods, we will 
 consider papers investigating a specific modality (e.g., vision or 
 sonar) with the aim of generalizing the findings to other modalities. 
 Also, although we are interested in submissions detecting patterns in 
 sensory data, we would especially like to encourage submissions 
 addressing how richer theories (such as entities, relations, and 
 causality) might be derived from sensor data.


 FORMAT

 This one-day workshop will begin with an explanation of the workshop's 
 focus and research overview. We will decompose the workshop into themes 
 that concern learning rich representations from sensor data: tasks, 
 techniques, evaluations, or demonstrations. We will include invited 
 talks from senior researchers who can summarize their long-term research 
 on this topic. We will also include one or more panels that focus on 
 the themes listed above, and their challenges.


 SUBMISSION

 Submissions are due by March 28th. You are invited to submit through 
 EasyChair (https://www.easychair.org/conferences/?conf=replearn2013). 
 All submissions should be in AAAI's 2-column format 
 (http://www.aaai.org/Publications/Author/author.php), and must not have 
 been published elsewhere. Research papers should not exceed 6 pages, 
 and position papers should not exceed 3 pages. All submissions will be 
 refereed based on their relevance, originality, significance and soundness.


 ORGANIZING COMMITTEE

 Marc Pickett (Naval Research Laboratory, Washington, DC, USA, 
 marc.pickett.ctr@nrl.navy.mil)
 Ben Kuipers (University of Michigan, Ann Arbor, MI, USA)
 Yann LeCun (New York University, New York, NY, USA)
 Clayton Morrison (University of Arizona, Tucson, AZ, USA)


 ADDITIONAL INFORMATION

 For additional information, please visit the supplemental workshop site at
 http://www.marcpickett.com/RepLearn2013/