There has been considerable progress in robotics in the last years allowing robots to successfully contribute to our society. We can find them from industrial environments, where they are nowadays established, to domestic places, where their presence is steadily rising.

This workshop explores the following question: "How well prepared are learning robots to be social actors in daily-life home environments in the near future?". This workshop is therefore not only an opportunity to address this focus on the latest scientific contributions on bio-inspired learning social robotics, but also links this with the presence of robots in people daily-life environment. Thus, the main goal of this workshop is offering a common space for roboticists from different fields of expertise to discuss the current state-of-the-art of learning methods in robotics specially applied to home scenarios and recent developments in assistive robots.

Roboticists are aware of the big challenges that involve working with service and assistive robots in home environments to develop real robot domestic applications. For instance, the RoboCup initiative founded a specific league "RoboCup@Home league" to aim the development of highly interactive intelligent robots to perform tasks in new and complex environments while being able to anticipate and resolve conflictual situations that may lead to mistakes or incomplete performance.

Such complex learning tasks in home environments can include among others learning to:

• Provide help in home services,
• Wipe/tidy up a table, floor, or room,
• Cook a meal,
• Be of assistance for elderly people,
• Be a conversational companion.

Each of these domestic activities have been mostly investigated and developed as more simple restricted tasks in controlled environments. However, the learning and further development of real complex tasks in actual dynamic scenarios is still an open issue in robotics. Intelligent robots operating around us should be able to know where it is located itself, detect people, learn and recognize faces, learn new objects, understand action-object opportunities, and furthermore they should learn to behave cooperatively in domestic scenarios.

In order to accomplish these complex domestic tasks successfully, robots and roboticists have to deal with many challenges such as perception, pattern recognition, navigation, and object manipulation, all of that in varying environmental conditions. Such challenges can only be addressed if the robot constantly acquires and learns new skills, either autonomously or from parent-like trainers.

This workshop principally targets bio-inspired developmental learning and psychologically motivated approaches for domestic environments. These approaches are inspired by how humans develop knowledge through interactions with their environment. Latest research has indicated important advances for developing learning approaches in domestic robotic applications. Indicative example approaches with special interest for this workshop include:

• Interactive reinforcement learning,
• Neural sequence learning
• Policy and reward shaping,
• Learning of object affordances and contextual affordances,
• Predictive learning from sensorimotor information,
• Learning understanding of environment ambiguity,
• Learning with hierarchical and deep neural architectures,
• Bootstrapping complex action learning in robots,
• Learning supported by external trainers by demonstration and imitation,
• Parental scaffolding as a bootstrapping method for learning.

In this context, learning can be understood as a perceptual or behavioral problem which can be addressed by bio-inspired methods and different forms of learning.