The Conference on Neural Information Processing Systems, abbreviated NeurIPS, is the largest annual gathering in machine learning. Its own about page states that “the conference was founded in 1987 and is now a multi-track interdisciplinary annual meeting that includes invited talks, demonstrations, symposia, and oral and poster presentations of refereed papers.” It is run by the Neural Information Processing Systems Foundation, a non-profit established to govern the meeting.
NeurIPS grew out of a late-1980s community working at the intersection of neuroscience, statistics, and the then-unfashionable field of neural networks. Over the following decades it became the premier venue for the deep-learning revolution: results such as AlexNet, generative adversarial networks, sequence-to-sequence learning, and word embeddings were all presented there. The conference’s growth tracked the field’s, swelling from a few hundred attendees to many thousands and adopting a lottery and registration cap to manage demand.
For most of its history the conference used the acronym NIPS, before officially switching to NeurIPS in 2018 after complaints that the old name was open to crude puns and created an unwelcoming environment. The Foundation also moved its web address from nips.cc to neurips.cc.
Why business readers should care: NeurIPS is where the techniques that now power commercial AI products were first peer-reviewed and debated, and the company affiliations of its accepted papers are a useful early signal of where applied capability is heading.