[D] How often do reviewers decrease their initial scores after rebuttal period ends in CVPR?

· Source: Machine Learning · Field: Science & Research — Research Methodology & Innovation, Mathematics & Computational Sciences · Depth: Intermediate, quick

Summary

Discussions among researchers on Reddit indicate that initial reviewer scores for papers submitted to major computer vision conferences like CVPR and ICCV can decrease after the rebuttal period, though it is not a frequent occurrence. One user reported experiencing a score downgrade once across 15 submissions or reviews. Another user's ICCV paper saw scores drop from 5,3,3 to 5,4,2, leading to rejection. An Area Chair (AC) noted that downgrades are relatively common for borderline initial ratings, as reviewers are encouraged to avoid such final scores. Evasive or combative rebuttals, or a reviewer convincing others of a major issue during discussion, are primary causes. ACs rarely introduce new issues post-rebuttal to avoid authors being unable to respond.

Key takeaway

For AI Scientists submitting to CVPR or ICCV, understand that initial reviewer scores are not final and can decrease after the rebuttal period. Focus on crafting clear, direct rebuttals that address reviewer concerns without being evasive or combative, especially if your paper has borderline initial ratings. Be prepared for the possibility of rejection even with seemingly good initial scores, and consider having a backup plan for resubmission.

Key insights

Reviewer scores can decrease post-rebuttal, especially for borderline ratings or due to ineffective rebuttals.

Principles

In practice

Topics

Best for: AI Scientist, AI Researcher, Research Scientist, AI Student

Related on AIssential

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine Learning.