IJCAI Related Work

Use this to audit novelty, prior work, and submission eligibility. Reopen the current CFP and submissions FAQ because resubmission and dual-submission rules are policy-sensitive.

Positioning checks

• Distinguish non-archival workshop or poster versions from formally published proceedings with DOI, ISBN, or ISSN.
• If the paper was rejected by another peer-reviewed conference in the last 12 months, prepare the required anonymized rejected version and reviews; the response cover letter may be optional depending on current rules.
• Cite author preprints and prior work in third person. A non-anonymous arXiv version may be allowed, but the IJCAI submission itself must remain anonymous.
• Do not submit a paper that is under review, accepted, or published at another formal archival venue during the IJCAI review period.
• If a substantially shorter extended abstract, poster, or position paper exists, declare it when current rules require and explain the new technical content.
• Make the related-work section decision-useful: closest competing claims, why they are insufficient, and exactly what is new here.

Positioning across IJCAI's breadth

Because IJCAI spans symbolic AI, search, planning, KR, constraint reasoning, multi-agent, game theory, ML, NLP, and vision, the closest prior work may live in a different subcommunity than the authors'. A reviewer from that adjacent area will know the literature you skipped.

Worked vignette: a search paper missing an adjacent baseline

A bidirectional-search paper cites only recent search venues and omits a closely related planning-community result. An IJCAI planning reviewer flags the gap as a novelty threat. Fix: add the planning result to the closest-works set, state precisely why it does not subsume the new bound, and write a one-sentence novelty contrast a non-search reviewer can verify. If the paper is a shortened version of a prior workshop poster, declare that and describe the added technical content.

Reviewer pushback and the venue-specific fix

• "Missed obvious prior art." Search the adjacent subfield, not just your home venue, since IJCAI reviewers cross subcommunities.
• "Looks like double submission." Confirm no concurrent archival submission and prepare the resubmission package if a recent rejection exists.
• "Related work is a list, not an argument." Convert it to closest claims, why insufficient, and what is new.

Output format

[Eligibility] clear / needs declaration / risky
[Closest works] <3-5 items and distinction>
[Resubmission package] <needed/not needed/unknown>
[Dual-submission risk] <none/issues>
[Novelty sentence] <IJCAI-ready contribution contrast>