Irony alert: If you’re going to write articles about recurrent neural networks, make sure they don’t, well, recur.
The journal Neural Networks has retracted a 2012 paper by a group of researchers from Spain for publishing what amounted to a repeat of a 2011 article in a different but closely related journal. Scientific publishing, alas, is not like Hollywood, where remakes of movies and TV shows is not only acceptable, it seems to be the only flavor producers are willing to taste.
The article was titled “Hopf Bifurcation Stability in Hopfield Neural Networks” — Hopf, for those keeping score at home, is Eberhard Hopf, a famous (and politically controversial) mathematician and founder of something called ergodic theory — and it came from scientists at the University of La Laguna, in the Canary Islands.
According to the notice:
This article has been retracted at the request of the Editors in Chief as it contains unacceptable overlap with earlier research published by R.L. Marichal, J.D. Pineiro, E.J. Gonzalez, and J. Torres in “Stability of Quasi-Periodic Orbits in Recurrent Neural Networks,” Neural Processing Letters Volume 31, Issue 3, Pages 269–281, June 2010. DOI http://dx.doi.org/10.1007/s11063-010-9138-9.
The paper has yet to be cited, according to Thomson Scientific’s Web of Knowledge.