Learning to Detect

Neev Samuel*, Tzvi Diskin, Ami Wiesel

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

293 Scopus citations


In this paper, we consider multiple-input-multiple-output detection using deep neural networks. We introduce two different deep architectures: a standard fully connected multi-layer network, and a detection network (DetNet), which is specifically designed for the task. The structure of DetNet is obtained by unfolding the iterations of a projected gradient descent algorithm into a network. We compare the accuracy and runtime complexity of the proposed approaches and achieve state-of-the-art performance while maintaining low computational requirements. Furthermore, we manage to train a single network to detect over an entire distribution of channels. Finally, we consider detection with soft outputs and show that the networks can easily be modified to produce soft decisions.

Original languageAmerican English
Article number8642915
Pages (from-to)2554-2564
Number of pages11
JournalIEEE Transactions on Signal Processing
Issue number10
StatePublished - 15 May 2019

Bibliographical note

Funding Information:
Manuscript received May 16, 2018; revised November 1, 2018 and January 17, 2019; accepted January 25, 2019. Date of publication February 15, 2019; date of current version April 12, 2019. The associate editor coordinating the review of this manuscript and approving it for publication was Prof. Marco Lops. This work was supported in part by the Heron Consortium and by ISF under Grant 1339/15. (Corresponding author: Neev Samuel.) The authors are with the School of Computer Science and Engineering, The Hebrew University of Jerusalem, Jerusalem 9190401, Israel (e-mail:, neev.samuel@mail.huji.ac.il; zvidiskin@gmail.com; amiw@cs.huji.ac.il). Digital Object Identifier 10.1109/TSP.2019.2899805

Publisher Copyright:
© 1991-2012 IEEE.


  • MIMO detection
  • deep learning
  • neural networks


Dive into the research topics of 'Learning to Detect'. Together they form a unique fingerprint.

Cite this