ICLR 2017

This is an old revision of the document!


6th International Conference on Learning Representations

(ICLR 2018)

Overview

The performance of machine learning methods is heavily dependent on the choice of data representation (or features) on which they are applied. The rapidly developing field of deep learning is concerned with questions surrounding how we can best learn meaningful and useful representations of data. We take a broad view of the field and include topics such as feature learning, metric learning, compositional modeling, structured prediction, reinforcement learning, and issues regarding large-scale learning and non-convex optimization. The range of domains to which these techniques apply is also very broad, from vision to speech recognition, text understanding, gaming, music, etc.

A non-exhaustive list of relevant topics:

  • unsupervised, semi-supervised, and supervised representation learning
  • representation learning for planning and reinforcement learning
  • metric learning and kernel learning
  • sparse coding and dimensionality expansion
  • hierarchical models
  • optimization for representation learning
  • learning representations of outputs or states
  • implementation issues, parallelization, software platforms, hardware
  • applications in vision, audio, speech, natural language processing, robotics, neuroscience, or any other field

The program will include keynote presentations from invited speakers, oral presentations, and posters.

ICLR features two tracks: a Conference Track and a Workshop Track. Submissions of extended abstracts to the Workshop Track will be accepted after the decision notifications for Conference Track submissions are sent. A future call for extended abstracts will provide more details on the Workshop Track.

Some of the submitted Conference Track papers that are not accepted to the conference proceedings will be invited for presentation in the Workshop Track.

What Is Changing from ICLR 2017

1) Submissions will be double blind, meaning that reviewers cannot see author names when performing reviews, and authors cannot see reviewer names. While we will still use Open Review to host papers and allow for public discussions that can be seen by all, comments that are posted remain anonymous.

2) If someone wants to cite a paper during the review period, OpenReview will provide a BibTeX entry that does not list the authors, but does give the title, year and URL. Only at the end of the review period will the authors be revealed.

3) We will have only one round of review process, where the reviewers will provide their assessment of the paper. There will still be a discussion period between the authors and reviewers after the initial review round.

ICLR Submission Instructions

By October 27 - 5:00pm EST, authors are asked to submit their paper to:

https://openreview.net/group?id=ICLR.cc/2018/Conference

The initial paper needs to be submitted by this date, however authors are encouraged to update their submission as desired. Note that the initial paper that will be reviewed is based on that submitted by the October 27th deadline. There is no strict limit on paper length. However, we strongly recommend keeping the paper at 8 pages, plus 1 page for the references and as many pages as needed in an appendix section (all in a single pdf). The appropriateness of using additional pages over the recommended length will be judged by reviewers. Authors are encouraged to participate in the public discussion of their paper, as well as any other paper submitted to the conference. Submissions and reviews are both anonymous. For detailed instructions about the format of the paper, please visit www.iclr.cc.

Dual Submission Policy

Submissions that are identical (or substantially similar) to versions that have been previously published, or accepted for publication, or that have been submitted in parallel to other conferences or journals are not allowed and violate our dual submission policy. However, papers that cite previous related work by the authors and papers that have appeared on non-peered reviewed websites (like arXiv) or that have been presented at workshops (i.e., venues that do not have a publication proceedings) do not violate the policy. The policy is enforced during the whole reviewing process period.

Withdrawing Policy

Authors have the right to withdraw papers at any time until paper notification. However, once a paper has been submitted to ICLR, a withdrawn paper will be moved to a “withdrawn papers” section and thus will not get reviewed or receive a paper decision. However it will continue to get hosted by OpenReview and be publically visible.

Reviewing Process

  1. Submissions to ICLR are uploaded on OpenReview, which enables public discussion. All comments are publicly visible (even for people who are not logged in) but the author of a comment can decide to post anonymously or not. Log in is required before posting any comment.
  2. Authors are encouraged to revise their paper as many times as needed up until the paper deadline of October 27. Note authors can participate in the discussion about their paper as well as any other paper submitted to the conference, at any time.
  3. By November 27 2017, reviewers need to post their full review. Reviews are anonymous and publicly visible in Open Review. Once the reviews are posted, authors are free to make changes to their paper as often as needed.
  4. The rebuttal period will be from November 27, 2017 until January 5, 2018, where authors can address reviewer comments and make changes to the paper. During this period, any submission that is cited will be given an anonymous BibTex entry.
  5. On January 29, 2018, authors will be notified about the acceptance or rejection of their paper. Some of the rejected papers that distinguish themselves for their originality will be invited for presentation under the Workshop Track. After notification, all BibTex entries will be de-anonymized.
  6. Papers that are not accepted to the Conference Track will be considered non-archival, and may be submitted elsewhere (modified or not), although the OpenReview site will maintain the reviews, the comments, and links to the versions submitted to ICLR.

General Chairs

Yoshua Bengio, Université de Montreal

Yann LeCun, New York University and Facebook

Senior Program Chair

Tara Sainath, Google

Program Chairs

Iain Murray, University of Edinburgh

Marc’Aurelio Ranzato, Facebook

Oriol Vinyals, Google DeepMind

Steering Committee

Aaron Courville, Université de Montreal

Hugo Larochelle, Google

Contact

The organizers can be contacted at iclr2018.programchairs@gmail.com