NeurIPS / ICML / ICLR

The three most influential annual academic conferences in machine learning.

NeurIPS (Conference on Neural Information Processing Systems, called NIPS until 2018, usually held in December), ICML (International Conference on Machine Learning, usually held in summer), and ICLR (International Conference on Learning Representations, launched in 2013 by Yoshua Bengio, Yann LeCun, and others, with public review on OpenReview) are together known as machine learning's “big three” conferences. In computer science, conference papers are the primary form of publication, and these three have low acceptance rates and high impact; China's CCF ranking classifies them as top-tier “CCF-A” venues. Methods-heavy embodied-AI work — VLA models, world models, reinforcement-learning algorithms — is often submitted to these three, while more systems- and real-robot-focused work tends to go to CoRL, RSS, or ICRA instead.

ExampleThe Decision Transformer paper was published at NeurIPS 2021; many VLA and world-model papers are submitted to ICLR.

Also called

Machine Learning's “Big Three” Conferences

Related

Top-tier Conferences & Journals (CCF-A Venues) · Conference on Robot Learning · Robotics: Science and Systems · IEEE International Conference on Robotics and Automation · CVPR / ICCV / ECCV · arXiv Preprint

Sources

NeurIPS
ICML
ICLR