In recent years, there has been a lot of interest in algorithms that learn feature representations from unlabeled data. Deep learning algorithms such as deep belief networks, sparse coding-based methods, autoencoder variants, convolutional networks, ICA methods, and deep Boltzmann machines have shown promise and have already been successfully applied to a variety of tasks in computer vision, audio processing, natural language processing, information retrieval, and robotics. In this workshop, we will bring together researchers who are interested in deep learning and unsupervised feature learning, review the recent technical progress, discuss the challenges, and identify promising future research directions.

The workshop invites paper submissions that will be either presented as oral or in poster format. Through invited talks, panel discussions and presentations by the participants, this workshop attempts to address some of the more controversial topics in deep learning today, such as what is a good representation, how it could be learned, and what obstacles need to be addressed in future research. Panel discussions will be led by the members of the organizing committee as well as by prominent representatives of the vision and neuroscience communities.

The goal of this workshop is two-fold. First, we want to identify the next big challenges and propose research directions for the deep learning community. Second, we want to bridge the gap between researchers working on different (but related) fields, to leverage their expertise, and to encourage the exchange of ideas with all the other members of the NIPS community.


We solicit submissions of unpublished research papers. Papers should be at most 8 pages (plus 1 additional page containing references only) and must satisfy the formatting instructions of the NIPS 2012 call for papers. Style files are available at http://nips.cc/PaperInformation/StyleFiles. Please note that the reviewing is double blind, so your manuscript should not contain authors’ identifying information. Papers should be submitted through https://cmt.research.microsoft.com/DL2012/ no later than 23:59 EST on Sunday, September 16, 2012.
We encourage submissions on the following and related topics:

unsupervised feature learning algorithms
deep learning algorithms
semi-supervised and transfer learning algorithms
inference and optimization
theoretical foundations of unsupervised learning
theoretical foundations of deep learning
applications of deep learning and unsupervised feature learning

The best papers will be awarded by an oral presentation, all other papers will have a poster presentation accompanied by a short spotlight presentation.