ICML 2026 reviews spark borderline anxiety

A 3.5 average on the new 1-6 scale places the paper in a high-uncertainty zone where reviewers are explicitly instructed to use borderline ratings sparingly. Success depends entirely on the rebuttal's capacity to address specific empirical demands, such as parameter sweeps, to flip "Borderline Rejects" into accepts. The transition from theoretical physics often brings novel perspectives but can struggle with the rigid empirical conventions of top-tier computer science conferences. Community sentiment suggests that "raiseable" scores put immense psychological pressure on authors to deliver a perfect final response. With notification dates set for late April, the 4433 vector represents the most stressful outcome for researchers awaiting conference funding.