
[Home](../index.html) [Speakers](../speakers.html) [Program](../program.html) [Call for Papers](../call.html) [Accepted Papers](../papers.html) [Organizers](../organizers.html)
 [Emre Ugur](https://colors.cmpe.boun.edu.tr/) is an Associate Professor in Dept. of Computer
Engineering, Bogazici University, the chair of the Cognitive Science
MA Program, the vice-chair of the Dept. of Computer Engineering, and
the head of the Cognition, Learning and Robotics (CoLoRs) lab. He received his BS, MSc, and Ph.D.
degrees in Computer Engineering from Middle East Technical University
(METU, Turkey). He was a research assistant in KOVAN Lab. METU
(2003-2009); worked as a research scientist at ATR, Japan (2009-2013);
visited Osaka University as a specially appointed Assist.&Assoc.
Professor (2015 & 2016); and worked as a senior researcher at the
University of Innsbruck (2013-2016). He is interested in robotics,
robot learning, and cognitive robotics.
**Title:** Safe robot learning and control with Conditional Neural Processes
**Abstract:** The robots learning and acting in environment interactions
need to take into account the physical and social constraints of the
environments and the tasks. In the first half of this talk, I will
present our work ACNMP, namely Adaptive Conditional Neural Movement
Primitives, that allows safe and efficient policy improvement in novel
environments. Following a learning from demonstration phase, our model
enables policy improvement by simultaneous training of our model with
supervised learning (SL) Reinforcement Learning (RL). ACNMP enables
the system to safely extrapolate to situations where pure LfD and RL
fail. In the second part of my talk, I will present our learning
social navigation framework, which learns global and local social
controllers of the mobile robot from observations. We leverage a
state-of-the-art, deep prediction mechanism to detect situations
not similar to the trained ones, where reactive controllers step in to
ensure safe navigation.