RAM Special Issue on Deep Learning and Machine Learning in Robotics-Final manuscripts uploaded by authors

On 10 Mar, 2020

At IEEE 445 Hoes Lane, Piscataway Township, NJ 08854, USA

Posted by alexis.simoes@ieee.org

Categories: Deadlines

Rate:
Save to calendar


Call for Papers
IEEE Robotics & Automation Magazine Special Issue on
“Deep Learning and Machine Learning for Real World Scenarios in Robotics and Automation”
Deadline for paper submission: 1 August 2019
Publication Schedule: June 2020
First decision: 1 November 2019
Final decision: 20 February 2020
Deep learning and Machine Learning have gone through a massive growth in the past several years. In many domains, such as perception, vision, image recognition, image captioning, speech recognition, machine translation, and board games, in particular, deep learning has drastically outperformed traditional methods and overtaken them to become the method of choice. Will the same happen to robotics and automation? These approaches typically require massive amounts of labeled data, i.e., big data, and massive amounts of compute. In many real robotics and automation applications data is abundant but labeling sparse and expensive. (Deep) reinforcement learning often requires significantly more iterations than are feasible on real systems. Hence collecting sufficient amounts of data is impractical at best. Therefore, a lot of work is done in purely digital or virtual environments. In this special issue we will focus on approaches that have been validated on real world robots, scenarios, and automation problems. While a lot of progress has been achieved on this front in robotic and automation applications, still a lot of progress needs to be made in order to render deep learning approaches directly applicable. Robots and automation systems are interacting with the real world. Hence mistakes that might be costly in terms of lost revenue in approaches that operate in a purely digital world, can cause significant damage and loss of human lives. Therefore, safe learning becomes paramount. A related issue is interpretable learning, i.e. the capability to interpret learning processes, moving towards approaches where humans have the option to be in control and understand with sufficient human-readable details the decision processes of the machine. Successful applications in ‘neighboring’ fields characterized by limited amounts of sparse, labeled data coming from physical systems will also be considered.
Papers should follow the standard RAM guidelines. A full peer-review process will be utilized to select papers for the special issue. Submissions should be made through the RAM submission website by August 1, 2019.
Contributions are expected to present original research on deep learning and machine learning with real world applications in robotics and automation.
The topics of interest include but are not limited to:
deep learning
supervised
unsupervised
reinforcement
sample efficient learning
new methods
use of models
simulation to real transfer
data augmentation
embedding prior knowledge

safe learning
confidence estimates
guarantees
verification
interpretable learning

real applications and use case scenarios of deep/machine learning
robotics
perception
control
planning
navigation
manipulation and grasping

automation
maintenance and inspection
production
quality management and assurance
product tracking

success stories of deep/machine learning technologies in robotics and automation
common issues and solutions in deep/machine learning applications in robotics and automation and neighboring fields such as:
gravitational waves detection
geophysics
high energy physics

Guest editors
Fabio Bonsignorio, Heron Robots
David Hsu, National University of Singapore
Matthew Johnson-Roberson, University of Michigan
Jens Kober, TU Delft


2020-03-10 00:00:00
-0001-11-30 00:00:00

Easy Links