This
document contains information about the process of submitting a paper to
NIPS*2007. You can
also find the Call
for Papers, evaluation
criteria for NIPS papers, a page with the
style
files, and a complete list of
keywords.
On September 5, the
accept/reject decisions will be sent.
NIPS accepts only electronic submissions at:
http://nips2007.confmaster.net starting Saturday, May 26.
The submission
deadline is 11:59PM UTC (7:59PM Eastern Daylight, 4:59PM Pacific Daylight) on
Friday, June 8, 2007. The server may be open for a few more minutes after the deadline to allow for
possible server load, so if you find the server closed, you have definitely
missed the submission deadline.
There will be absolutely no extensions given to the submission deadline.
All submissions
must be in pdf format. As we do not
accept paper submissions, particular care should be taken to ensure that your
paper prints well on the reviewers' hardware. Please consult Section 6 in the
PDF/PS example file. Note that the
maximal file length for submissions is 10MB.
You can upload an
early version of your paper well before the deadline. If you start registering
your paper only a few minutes before the deadline, you may not have enough time
to fill in all the forms. Replacing an earlier version later is no problem and
does not take as long ("upload early and often").
We also wish to draw your attention to the newly
formulated NIPS evaluation
criteria, which will form the basis for this year's acceptance decisions.
If your paper is accepted, there
will be an opportunity after the meeting to revise accepted
manuscripts.
If you are not only
an author but also a reviewer, please use the same login. If your paper is
registered by your co-author, tell them which of your email addresses to use
(the same as the one you are using as a reviewer).
Your submission must follow the style files that are available at this location.
Papers departing
from the formatting guidelines and all papers longer than 8 pages will be
rejected without review.
The NIPS reviewing
process is now double blind, so please follow the style files provided and make
sure the submission does not include author names and affiliations. In
particular, here are some guidelines for submitting double-blind
papers:
·
Do
not include any author name or affiliation in the title.
·
Do
not include any acknowledgments.
·
Refer
to your own published work in the third person.
That is, use “In the previous work of Jones et al. [2]”, not “In our
previous work [2]”.
·
If
you cite your other papers that are not widely available (e.g., a
difficult-to-find technical report or a paper currently under review), use
anonymous author names in the citation, e.g., an author of the form “A.
Anonymous”.
Papers not
satisfying these guidelines will not be reviewed.
The purpose of the
keywords is to assist assignment of papers to area chairs and
reviewers. However, the separation between different
areas does not play a major role anymore, and authors should refrain from
choosing keywords based on past rumors of what may increase their chances of
acceptance. The keywords this year are
hierarchically-structured, with each top-level area elaborated by more specific
topics. If a low-level keyword (e.g.,
Gaussian processes) describes your paper well, use that; there is no need to
also choose the top-level category to which it belongs (e.g., Probabilistic
Models). Of course, you should still use
all the keywords that are relevant to your paper (e.g., Gaussian Processes and
Robotics-Perception). Note that the same
keyword (e.g., Learning) can appear in multiple categories; these are not the
same keyword. The list of keywords to
choose from can be found here.
As in last year's
process, because of double-blind reviewing, authors can upload supporting
material with their paper. Please make sure that the supporting material is
also anonymized, removing your name from any extra
material, and referring to previous work in the third person.
Such extra material
may include long technical proofs that do not fit into the paper, image, audio
or video sample outputs from your algorithm, animations that describe your
algorithm, details of experimental results, or even source code for running
experiments.
Note that the
reviewers and the program committee reserve the right to judge the paper solely
on the basis of the 8 pages of the paper: looking at any extra material is up
to the discretion of the reviewers and is not required.
To submit a paper
with extra material, please create an archive file and upload it to ConfMaster. This archive file can be in the .zip format
(using WinZip or equivalent), or in .tar.gz format
(using tar and gzip). There is a strict upper limit
of 10 megabytes for any submission, including both the paper and the extra
material. Archives larger than this will
be rejected by ConfMaster. Please be
circumspect, and upload only files
that are directly useful for evaluating your paper.
To reduce load on the server, we request that large files be uploaded
not right at the time of the deadline. Please upload material that is
ready early (e.g., videos or images) as soon as possible.
If you do create an
archive, please name your paper file in the following way: start with “paper_”,
then use a few words from your title to disambiguate
your paper from others, then end with “.pdf” or “.ps” as appropriate. For extra files, start with a
description of the extra material, such as “proof_” , or “appendix_”, or
“video_”; then use the same words as your main submission, then the appropriate
extension.
Please use these
naming schemes: failure to do so will break our automated scripts and result in
your material not being available to the reviewers.
Submissions that
are identical (or nearly identical) to versions that have been previously
published or that have been submitted in parallel to other conferences whose
audience significantly overlaps with the one of NIPS are not appropriate for
NIPS. Exceptions to this rule are the following:
(1) Shorter
write-ups of longer papers that have been recently (i.e. in the current
calendar year) submitted to journals.
(2) Papers whose
content is currently under review elsewhere or has very recently been
published, but only in venues that are particularly inaccessible to the NIPS
audience. In these cases, the NIPS submission should
involve a substantial revision of the original paper, in a way that specifically
highlights and expands on the relevance of the work to the NIPS
community. The differential contribution between the
original paper and the NIPS submission will be a factor in the
decision.
Authors of such papers must anonymously cite
the earlier work, and include an anonymized copy in
the supplementary materials.
Examples of
conferences that are too close to NIPS and where double submission will not be
considered: ICML, UAI, COLT, ICCV, ECCV, CVPR, ECML, AI&STATS, KDD, ICANN,
IJCNN, WCNN, SODA, FOCS, STOC, ACL, EMNLP.
Examples of
conferences where the papers are generally unfamiliar to the NIPS audience
include application-oriented conferences, such as SIGGRAPH, SIGIR, RECOMB/ISMB,
or WWW.
Starting July 18, authors
will be given the chance to see the reviews and respond to their content.
Giving feedback is
optional: your feedback can be used by the reviewers to alter their judgment
about the paper. The system allows you to respond to each review individually in
a text box. These responses are due back by 23:59:59 UTC on Monday, July 23.
A few things to
note about the author response:
§
Due
to a limitation in ConfMaster, you can submit
feedback only once per review. We strongly recommend that you read the reviews,
discuss the response amongst your co-authors, draft the response off-line, then upload the response(s). Only the contact author can
upload the response.
§
Note: the system has only accepted your response if you
can read it after you click “Submit”. Otherwise, it is not registered in the
system and reviewers cannot read it.
§
We've
limited each response to be 250 words long (per review). If you exceed 250
words according to the algorithm, the system will silently reject the response
and allow you to try again.
§
The
word count is computed somewhat differently than the UNIX wc program: a word is a contiguous string of one or
more characters from the set A-Z, a-z, underscore, or backslash. Digits are
completely ignored. If you have Perl
installed on your computer, you can emulate ConfMaster's
word counting program by saving your response in a file, and typing
perl -ne 'END
§
Remember
that reviewing is double-blind this year. Do not include any information in the
response that can identify you or your co-authors. Please do not include any
URLs in your response.
§
Every
response to a paper can be seen by every reviewer, so there is no need to
repeat information between responses.
Finally, we
recommend using the response to influence the reviewers where they can be
influenced. If a reviewer has expressed
uncertainty about an issue, or is making an incorrect assumption, or has
misunderstood a point in the paper, that is ideal to address in an author's
response. It may be difficult to change
a reviewer's value judgment about a paper, such as overall clarity or
significance. There is no need to
respond to every minor question or suggestion for improvement. In any
event, the best strategy is to be
polite and professional.
Quick links:
Home
Call for Papers
Author and submission instructions
Style files
Paper submission site
Paper evaluation criteria
Reviewer instructions
FAQ
Questions?
Comments? Please send email to nips07 AT cs.stanford.edu (non-standard form used
to prevent spam).