Advanced Learning Techniques for Autonomous Robots in the Internet of Things (ALTARIoT)

In the past twenty years, Machine Learning (ML) has enabled a number of key technologies that have revolutionized many aspects of our daily lives. Notable examples include spam filtering, automated fraud detection, face recognition, predictive medicine. 

In addition, due to the exponential growing of interests and applications for IoT, ML-based systems have evolved to the point of being able to make sense of complex and huge sets of data collected by IoT devices, and derive meaningful decisions: large-scale recommender systems provide buying suggestions to online shoppers, and self-driving vehicles can algorithmically predict whether a pedestrian will cross and stop if required.
Furthermore, because of their inherent ability to deal with complex information, ML techniques find a natural application in the creation of autonomous robotic/multi-agent systems. The overarching hypothesis is that ML will facilitate the creation of robotics systems that can autonomously operate in complex environments by exploiting data collected by IoT devices, adapt to changing circumstances and predict/avoid dangerous situations.
This special session will solicit contributions that identify the challenges related to the application of ML (particularly Deep learning) techniques to robotic/multi-agent systems fully connected as IoT devices and propose novel methods to enhance the autonomous capabilities of robots and agents.