Open Internet by MindsNet

Inconsistent Paper Scoring in ICML Review Process

The ICML 2026 review process exhibits significant variance in paper scores across different batches, with some batches having very few papers scoring above 3.5 and others averaging around 3.75. This inconsistency raises concerns about fairness and reliability in the evaluation process. The cause of this variance is unclear, but potential factors include differences in reviewer harshness and domain-specific evaluations. The inconsistency may impact the acceptance and perception of papers across different batches.

Computing & Technology, Computer Science, Machine Learning