ICML rebuttal favors citations, scope, AC framing
The post asks how to respond in an ICML rebuttal when reviewers keep claiming a method is not novel, despite the authors having addressed other concerns. The author argues the work is mostly novel, produces unexpectedly strong results, and combines existing components in a new domain while also introducing new components, but the reviewers are dismissing those points. Commenters suggest asking for sources, clarifying that novelty can come from a new domain/application and composition, and focusing the rebuttal on persuading the area chair rather than trying to “win” over a fixed reviewer.
Hot take: this is less about proving the reviewer wrong and more about making the novelty claim hard to ignore at the AC level.
- –If the contribution is a novel composition in a new domain, say that plainly and separate it from claims about individual components.
- –Add citations or a concise “to the best of our knowledge” novelty statement; if the reviewer alleges prior art, ask them to point to it.
- –Distinguish between methodological novelty and empirical surprise: strong results alone do not prove novelty, but they can support the significance of the contribution.
- –Aim the rebuttal at the AC with a calm, evidence-based framing instead of debating tone or intent with the reviewer.
DISCOVERED
52d ago
2026-04-05
PUBLISHED
52d ago
2026-04-05
RELEVANCE
AUTHOR
Derpirium