López Morales, Manuel JoséChen Hu, KunGarcía-Armada, Ana2020-06-152021-03-172020-11-1012th IEEE/IET International Symposium on Communication Systems, Networks and Digital Signal Processing, (CSNDSP), 20-22, July 2020, Porto, Portugal (Online), 4 pp.978-1-7281-6743-5https://hdl.handle.net/10016/30612Proceeding of: 12th IEEE/IET International Symposium on Communication Systems, Networks and Digital Signal Processing, (CSNDSP), 20-22, July 2020, (Online).The performance of communication systems dependson the choice of constellations, designed in an end-to-endmanner. In case of a mathematical intractability, either becauseof complexity or even lack of channel model only sub-optimal solutionscan be provided with an analytical approach. We presentend-to-end learning, a recent technique in communications tolearn optimal transmitter and receiver architectures based ondeep neural networks (DNNs) architectures. We discuss cases inwhich this technique has been used to design constellations inwhich channel model intractability repressed from a mathematicalanalysis.4eng© 2020 IEEE.Machine learningEnd-to-end learningChannel imperfectionsConstellation designDeep neural networks architecturesconference paperTelecomunicacioneshttps://doi.org/10.1109/CSNDSP49049.2020.9249528open access14CC/0000030585