ICLR 2027 — International Conference on Learning Representations

ICLR 2027, the International Conference on Learning Representations, runs 26-30 April 2027 at the Moscone Center in San Francisco, California. ICLR is one of the three venues where foundational deep-learning work is first published, alongside NeurIPS and ICML, and it is the only one that reviews entirely in public on OpenReview: every submission, every review and every author rebuttal is readable by anyone. The organisers describe it as a five-day multi-track event mixing invited talks, oral paper presentations and poster sessions, with the main conference on 26-28 April and the final two days, 29-30 April, given over to workshops. Review is double blind. Submissions for the 2027 edition close on 25 September 2026, reviews are released on 5 November 2026, and final decisions land on 16 December 2026 — so by the time the conference opens, the accepted programme has been public for four months.

• Abstract deadlineSep 18, 2026
• Paper submission deadlineSep 25, 2026
• Reviews released to authorsNov 5, 2026
• Author-reviewer discussion closesNov 18, 2026
• Final decisionsDec 16, 2026

Date

April 26–30, 2027

Location

Moscone Center, San Francisco, California

Format

In-Person

Event Type

Conference

Status

Upcoming

Pricing

Paid

Organizer

International Conference on Learning Representations (ICLR)

Added

Aug 25, 2026

Updated

Aug 25, 2026

• ✓Machine learning researchers
• ✓AI/ML engineers
• ✓Applied research scientists
• ✓Heads of AI research
• ✓CTOs and VPs of Engineering at AI-first companies
• ✓PhD students and postdoctoral researchers

For a CTO or head of AI, ICLR is where the techniques that reach production in two years get argued out first — and because ICLR 2027 reviews openly on OpenReview, you can read that argument, including the objections, before a vendor sells you the product built on it. Attending puts research and applied-ML staff in front of the authors of methods your platform team is about to adopt, in the same metro as most frontier labs, and the poster sessions remain the cheapest recruiting channel in the field. The two workshop days are where evaluation, safety and infrastructure practice actually gets set.

• ✓Main conference 26-28 April: invited talks, oral paper presentations and poster sessions
• ✓Workshops on 29-30 April at the Moscone Center
• ✓Call-for-papers subject area: representation learning (unsupervised, self-supervised, semi-supervised and supervised)
• ✓Call-for-papers subject area: reinforcement learning
• ✓Call-for-papers subject area: probabilistic and generative models, causal reasoning and optimization
• ✓Call-for-papers subject area: fairness, safety and privacy
• ✓Call-for-papers subject area: datasets, benchmarks and infrastructure
• ✓Call-for-papers subject area: neurosymbolic and hybrid AI systems
• ✓Call-for-papers subject area: applications to robotics, neuroscience and the physical sciences
• ✓New ICLR 2027 policies on AI use, co-authorship and reciprocal reviewing in the author guidelines

• ★Fully public peer review on OpenReview — every submission, review, rebuttal and meta-review for ICLR 2027 is readable by anyone, which is not true of most AI conferences.
• ★One of the three venues (with NeurIPS and ICML) where foundational deep-learning results are first published, so the programme is a leading indicator of what reaches production one to two years later.
• ★Double-blind review: reviewers cannot see author names and authors cannot see reviewer names.
• ★Two dedicated workshop days on 29-30 April follow the three-day main conference, rather than being squeezed alongside it.
• ★Held in San Francisco in 2027, inside the same metro area as most frontier-model labs.
• ★ICLR 2027 introduces new author-guideline policies on AI use, co-authorship and reciprocal reviewing.

ICLR has not published 2027 registration rates. The official ICLR 2027 page states that venue and housing information will be posted there and that the official housing portal opens six months before the conference (from late October 2026). No pass tiers, student rate or early-bird deadline is available yet — the only dated commitments published so far are the submission and review milestones.

When is ICLR 2027?

ICLR 2027 takes place April 26–30, 2027 in San Francisco, California.

Where is ICLR 2027 held?

ICLR 2027 is held in San Francisco, California. The event format is in-person.

What is the ICLR 2027 submission deadline?

The next ICLR 2027 deadline is abstract deadline on Sep 18, 2026. Remaining dates: Paper submission deadline — Sep 25, 2026; Reviews released to authors — Nov 5, 2026; Author-reviewer discussion closes — Nov 18, 2026; Final decisions — Dec 16, 2026.

How much does ICLR 2027 cost?

ICLR 2027 is paid. ICLR has not published 2027 registration rates. The official ICLR 2027 page states that venue and housing information will be posted there and that the official housing portal opens six months before the conference (from late October 2026). No pass tiers, student rate or early-bird deadline is available yet — the only dated commitments published so far are the submission and review milestones.

Who should attend ICLR 2027?

ICLR 2027 is aimed at Machine learning researchers, AI/ML engineers, Applied research scientists, Heads of AI research, CTOs and VPs of Engineering at AI-first companies, PhD students and postdoctoral researchers.

What topics does ICLR 2027 cover?

ICLR 2027 covers Representation learning, Reinforcement learning, Probabilistic and generative models, Learning theory, AI safety, fairness and privacy, Graph and geometric learning, Neurosymbolic and hybrid AI, Datasets, benchmarks and infrastructure, Transfer, meta and lifelong learning, Interpretability of learned representations.

This page was written from 4 sources, 1 on domains other than iclr.cc.

1. 1.iclr.cc — 2027vendor
2. 2.iclr.cc — Datesvendor
3. 3.iclr.cc — CallForPapersvendor
4. 4.conferencedeadlines.com — iclr