CLOSED Call for Papers: Special Issue on Non-Euclidean Deep Learning

Paper submission due: CLOSED

First Notification: 1 January 2020

Revision: 1 March 2020

Final Decision: 1 May 2020

Publication: August 2020 (tentative)

Aim and Scope

Over the past decade, deep learning has had a revolutionary impact on a broad range of fields, from computer vision to natural language processing and speech analysis. Deep learning technologies are estimated to have added billions in business value, created new markets, and transformed entire industrial segments. Most of today’s successful deep learning methods such as convolutional neural networks (CNNs) rely on classical signal processing models that limit their applicability to data with underlying Euclidean grid-like structure, e.g. images or acoustic signals. Yet, many applications deal with non-Euclidean (graph- or manifold-structured) data such as social networks in computational sociology, molecular graphs in chemistry, interactomes in system biology, and 3D point clouds in computer vision and graphics. Until recently, the lack of deep learning models capable of correctly dealing with non-Euclidean data has been a major obstacle in these fields.

This special issue addresses the need to bring together leading efforts in non-Euclidean deep learning across all communities. We seek theoretical, algorithmic, and methodological advances, as well as new applications and uses. In particular, we are interested in works on the theoretical foundations of non-Euclidean deep learning.

Topics and Guidelines

This special issue targets researchers and practitioners from both industry and academia to provide a forum in which to publish recent state-of-the-art achievements in Non-Euclidean Deep Learning. Topics of interest include, but are not limited to:

• Theoretical aspects of non-Euclidean deep learning
• Relaxation of NP-hard problems using graph deep learning
• Generalized filters and pooling operators for non-Euclidean data
• Generative models (such as auto-encoders and GANs) for graphs and manifolds.
• Unsupervised learning on graphs
• Adversarial attacks on graphs, and protection against adversarial attacks
• Applications such as computer vision, graphics, shape analysis, biology, medical imaging, physics, computational social sciences, and complex network analysis

Deep learning on non-Euclidean domains is a common theme. This will be a criterion in evaluating submissions.

Before submitting your manuscript, please ensure you have carefully read the Instructions for Authors for IEEE Transactions on Pattern Analysis and Machine Intelligence. The complete manuscript should be submitted through TPAMI’s submission system (https://mc.manuscriptcentral.com/tpami-cs). To ensure that you submit to the correct special issue, please select the appropriate section in the drop-down menu upon submission. In your cover letter, please also clearly mention the title of the SI.

We are happy to receive extensions of works presented in top conferences but with a substantial revision (30 percent is generally considered "substantial"). Please visit www.computer.org/csdl/journal/tp/write-for-us/15083 for more information.

Guest Editors (alphabetical)

Michael Bronstein*, Imperial College London (UK), m.bronstein@imperial.ac.uk

Joan Bruna, New York University (USA), bruna@cims.nyu.edu

Taco Cohen, Qualcomm AI Research (Netherlands), tacos@qti.qualcomm.com

Marco Gori, University of Siena (Italy), marco@diism.unisi.it

Pietro Lio’, University of Cambridge (UK), pl219@cam.ac.uk

Jure Leskovec, Stanford University (USA), jure@cs.stanford.edu

Le Song, Georgia Institute of Technology (USA), lsong@cc.gatech.edu

Oriol Vinyals, DeepMind (UK), vinyals@google.com

Stefanos Zafeiriou*, Imperial College London (UK), s.zafeiriou@imperial.ac.uk

*primary contacts