Connectionists: IJCNN Special Session in Deep Reinforcement Learning

[Abdulrahman Altahhan]

Dear Colleagues we would like to cordially invite you to submit to the following special session in the coming IJCNN 2017 conference



IEEE International Joint Conference on Neural Networks (IJCNN 2017<http://www.ijcnn.org/paper-submission>)

Special Session on:                      Deep and Reinforcement Learning (DRL)
RL does not lend itself naturally to deep learning from an MDP framework perspective, and currently there is no uniformed approach to combine deep learning with realistic reinforcement learning despite very successful applications in virtual and gaming environments. Examples of important open questions are:

-          How to make the state-action learning process deep?

-          How to make the architecture of an RL system appropriate to deep learning without compromising the interactivity of the system?

-          How to make the learning suitable for a realistic scenario?

-          How to infuse a self-reflective process in the system?
Although recently there have been important advances in dealing with these issues, they are still scattered and with no overarching framework that promote them in a well-defined and natural way. This special session will provide a unique platform for researchers from Deep Learning and Reinforcement Learning communities to share their research experience towards a uniformed Deep Reinforcement Learning (DRL) framework in order to allow this important interdisciplinary branch to take-off on solid grounds. It will touch on the potential benefits of the different approaches to combine RL and DL. The aim is to bring more focus onto the potential of infusing reinforcement learning framework with deep learning capabilities that could allow it to deal more effectively with present applications such as realistic robotics, online streamed data processing that involves actions. Contribution is invited from all deep learning, reinforcement learning and deep reinforcement learning research.
Scope and Topics

-          Novel RL Techniques suitable for physical systems

-          Novel Deep and/or RL Algorithms

-          Novel Deep and RL Neural Architectures

-          Adaptation of existing RL Techniques for Deep Learning

-          Optimization and convergence proofs for DRL algorithms

-          Deeply Hierarchical RL

-          Deep and/or RL architecture and algorithms for Control

-          Deep and/or RL architecture and algorithms for Robotics

-          Deep and/or RL architecture and algorithms for Time Series

-          Deep and/or RL architecture and algorithms for Big Streamed Data Processing

-          Deep and/or RL architecture and algorithms for Optimizing Governmental Policy

-          Other Deep and/or RL theory and application...

Important Dates

-          Paper submission: November 15, 2016

-          Paper decision notification: January 20, 2017

-          Camera-ready submission: February 20, 2017

Organizers

Abdulrahman Altahhan(ab8556 at coventry.ac.uk<mailto:ab8556 at coventry.ac.uk>), COVENTRY UNIVERSITY, UK.

Vasile Palade, COVENTRY UNIVERSITY, UK.

Roozbeh Razavi-Far, University of Windsor, Canada.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/connectionists/attachments/20161014/be514d92/attachment.html>