Prehistory — From Neural Networks to the TransformerJames Bergstra et al. (Université de Montréal)

Bergstra et al. present Theano

Presented Theano at SciPy 2010 — a Python compiler for mathematical expressions that supports symbolic differentiation and GPU computation

Why it matters

The beginning of a setup in which frameworks determine the speed of research. Influenced the design of later deep learning libraries

Source — Primary sources first; secondary reports are labelled “Press”

Officialgithub.comhttps://github.com/Theano/Theano/blob/master/doc/citation.txt

Added in Week 35, 2026 · See that week

Nearby on the same axis

Rectified Linear Units Improve Restricted Boltzmann MachinesOfficialUnderstanding the difficulty of training deep feedforward neural networksPaperImageNet holds the first ILSVRCOfficialDeep Big Simple Neural Nets Excel on Handwritten Digit RecognitionPaper