ICLR 2026 oral with 2 rejects, 1 borderline reject

2026-03-19 · Source: Machine Learning · Field: Technology & Digital — Artificial Intelligence & Machine Learning · Depth: Intermediate, short

Summary

The ICLR 2026 review process is facing scrutiny due to several controversial Area Chair (AC) decisions, particularly concerning papers with highly divergent reviewer scores. One specific instance highlights a paper initially receiving scores of 8/4/2/2 (two rejects, one borderline reject) that was ultimately accepted as an oral presentation. This decision is unusual, as ACs typically expect a final score above 6, and reviewers rarely significantly update their scores. Conversely, other authors reported papers with strong initial scores, such as 8/6/4/4 and 8/6/6/6, being rejected. The lack of a discussion period this year, attributed to an OpenReview leak, forced ACs to make judgments on rebuttals without reviewer input, leading to perceived inconsistencies and concerns about ACs overriding reviewer consensus.

Key takeaway

For AI Scientists submitting to ICLR or similar conferences, understand that Area Chair decisions can significantly alter outcomes, even overriding strong reviewer consensus or elevating low-scored papers. You should be prepared for potential inconsistencies in the review process, especially when discussion periods are curtailed. Focus on clear rebuttals and ensure your work's practical implications are evident, as ACs may prioritize application potential, even for theoretically sound papers.

Key insights

Area Chair overrides and inconsistent review outcomes are raising concerns within the ICLR 2026 submission process.

Principles

• ACs can override reviewer scores.
• Reviewer scores rarely change significantly.
• Practicality can influence ICLR acceptance.

Method

The ICLR 2026 review process involved initial reviewer scores, author rebuttals, and AC decisions, with ACs making final judgments without reviewer discussion due to a platform leak.

In practice

• Scrutinize AC justifications for score changes.
• Consider theoretical contributions vs. practicality for ICLR.
• Document reviewer score updates carefully.

Topics

• ICLR Review Process
• Area Chair Overrides
• Paper Acceptance Criteria
• OpenReview Platform
• Peer Review Integrity

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

Related on AIssential

• [D] Saw this papaer from ICLR with scores 2,2,2,4 and got accepted, HOW — Area Chair overrides in peer review can lead to controversial paper acceptance/rejection decisions, especially with low reviewer scores.
• ICML 2026 Decision [D] — The ICML 2026 decision period is marked by high anxiety and frustration over the peer review process.
• AI-Assisted Peer Review at Scale: The AAAI-26 AI Review Pilot — AI systems can generate technically sound and preferred peer reviews at large conference scales.
• ICLR 2026 Response to Security Incident — An OpenReview API exploit led to a major ICLR security breach, compromising anonymity and academic integrity.
• Intelligence Is Not the Bottleneck: Validating an LLM First-Pass Manuscript Score Against Peer-Review Outcomes — An LLM-based system can accurately predict peer-review outcomes and provide structured feedback for academic manuscripts.
• Major conference catches illicit AI use — and rejects hundreds of papers — Watermarking can detect AI-generated content in peer review, leading to policy enforcement and paper rejections.

Open in AIssential →

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