Submission and publication

People encounter research through papers, code, data, blogs, and talks. Submission is an important form of communication with its own process and rules.

Distinguish the names

SCI / SCIE are citation indexes. Quartiles and impact factors are other evaluation information. Clarivate's overview explains index scope.

CCF A, B, and C are categories in the China Computer Federation's recommended list. How a university uses them for graduate admission or awards depends on its current rules. The list specifies its scope for full papers and other forms; see the official directory.

Top conference informally describes important venues such as NeurIPS, ICML, ICLR, CVPR, and ACL. Community attention, directory category, and institutional recognition have different bases.

Workshops organize discussion around a topic. Review, archival status, and later submission rules vary. Read the specific call; the ICLR 2026 organizer guide, for example, requires clarity on archival status.

Preprints and arXiv: arXiv shares papers but does not itself perform conference- or journal-style peer review. It also contains versions published elsewhere. Inspect version and publication details. See about arXiv.

Agree with collaborators before a first submission

Which community fits, what should the paper communicate, which evidence is missing, and how will contributions and authorship work? Then check that year's deadlines, page limits, template, anonymity, data, and code requirements.

Do not merely change the title in last year's template. Keep the current call open while writing and share changes with collaborators.

AI rules depend on year and role

Author use and reviewer use are different. ICML 2026's reviewing policy distinguishes prohibited and limited permitted use; reviewers follow the assigned policy. Author use and responsibility is addressed separately.

After reviews arrive

Separate unclear writing, genuinely insufficient evidence, and disagreement about value. Clarify with text and results; assess time and resources for added experiments; explain remaining disagreements specifically.

After rejection, revisit the question, experiments, and presentation. Read my experience in the freshman-year review, and methods on the writing page.

Resources for submission

Proceedings, OpenReview, and reading entries are in the research catalog. For writing and review preparation, see Huang and Ding's writing guide.