Origine: KDD 2009.
In writing your paper, we suggest you try to address the following questions, credited to George Heilmeier (for research projects):
- What are you trying to do? Articulate your objectives using absolutely no jargon.
- How is it done today, and what are the limits of current practice?
- What's new in your approach and why do you think it will be successful?
- Who cares?
- If you're successful, what difference will it make?
- What are the risks and the payoffs?
- What are the midterm and final "exams" to check for success?
In light of the above principles, we *suggest* the following guidelines for the paper content. Please exercise your discretion where required to make the paper as comprehensible as possible to the readers and reviewers. (The headings below are general categories and not necessarily meant to be used as section headings in your paper.)
Abstract
Try to include the following:- Motivation: one or two sentences on the problem and it significance;
- Results: a short paragraph on approach and results;
- Availability: a link to code, data, and supplementary materials, or a statement why this is not possible.
Motivation & Significance
What is the problem in layman terms and why is it important or significant?Problem Statement
Formal definition of the problem with any preliminary concepts.Prior Work & Limitations:=0D=0A What are the existing approaches, and their=
limitations?=0D=0A=0D=0A* Theory/Algorithm:=0D=0A - Discuss the main theoretica
l=
or algorithmic ideas of the paper;=0D=0A - Mention the main theorems, the=
intuition behind those, and their=0D=0A practical application. Move the=
proofs to the appendix, unless the=0D=0A proof itself is the main contributio
n;=
=0D=0A - Discuss your algorithmic solution at the conceptual level with=
=0D=0A pseudo-code, to convey the main ideas. Move minute (but=0D=0A =
practically important) implementation details to the appendix;=0D=0A -=
Discuss why you chose certain paths, and discuss unfruitful=0D=0A paths=
that you discarded. In other words, give both the=0D=0A theoretical (where=
appropriate) and algorithmic "insights" into=0D=0A your work.=0D=0A=0D=0A*=
Experiments:=0D=0A - Complete parameter settings and data descriptions=
should be=0D=0A provided (including any links to public resources);=0D=0A=
- Clearly specify the experimental procedure, including evaluation =0D=0A=
measures;=0D=0A - Compare to prior solutions, or at least to "strawman"=
solutions;=0D=0A - Clearly discuss the results and what they mean;=0D=0A=
- Only include the most relevant experiments here, using the=0D=0A appendix=
to provide any additional results (say on minor parameter=0D=0A tuning=
of your method, etc).=0D=0A=0D=0A* Discussion and Future Work:=0D=0A -=
Describe insights you gained, the limitations and applicability of=0D=0A=
Describe insights you gained, the limitations and applicability of=0D=0A=
your work, and directions for future research. Every solution has=0D=0A=
limitations, which should be explicitly mentioned.=0D=0A=0D=0A* References:=
=0D=0A Include the most relevant works, making sure all citations are complete=
=0D=0A (including editors, publishers, page numbers, etc.).=0D=0A=0D=0A*=
APPENDIX:=0D=0A You should use the appendix for supporting details. For=
example, you=0D=0A may use it to convey detailed technical/practical aspects=
of your=0D=0A implementation. You may use the appendix for theorem proofs,=
or for=0D=0A additional experimental results. Include include pointers=
in the=0D=0A main paper to relevant sections in the appendix.=0D=0A=0D=0A=
The appendix *is* an integral part of the paper, since it will provide=
=0D=0A details that are important for a proper appreciation of your work=
=0D=0A (e.g., for replicating or extending it, or for comparison). =0D=0A=
However, it should be possible on a first read-through to get a good =0D=0A=
understanding of the paper's contribution from the main part alone. =0D=0A=
Structuring the paper in this way provides a service to the reader, =0D=0A=
by separating main ideas from technical details. =0D=0A=0D=0A=0D=0A=0D=0A