The first International Conference on Learning Representations, ICLR, was held from May 2 to May 4, 2013, in Scottsdale, Arizona, co-located with the AISTATS conference. According to its archived program, the general chairs were Yoshua Bengio and Yann LeCun, two of the researchers who would later share the Turing Award for deep learning, with Aaron Courville, Rob Fergus, and Chris Manning as program chairs.
ICLR was created to give a fast-moving subfield its own home. The 2013 call for papers framed the topic around the observation that “the performance of machine learning methods is heavily dependent on the choice of data representation (or features) on which they are applied,” and invited work on “deep learning and feature learning, metric learning, kernel learning, compositional models, non-linear structured prediction, and issues regarding non-convex optimization.” This was the moment deep learning was becoming dominant, and ICLR quickly grew into one of the three top machine-learning conferences alongside NeurIPS and ICML.
ICLR is also known for helping pioneer open peer review in machine learning. It runs its reviewing on the OpenReview platform, where submissions, reviews, and author responses are visible to the public rather than kept confidential, an unusual degree of transparency for an academic venue.
Why business readers should care: ICLR’s founding marks the point at which deep learning had enough momentum to warrant its own institution, and its open-review model means the strengths and weaknesses of frontier methods are debated in public where anyone can read them.