Reconstructing a neural net from its output

By Charles Fefferman. 1994

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Philip M. Long
2. Peter L. Bartlett
3. Yoshua Bengio
4. J. Urgen Schmidhuber
5. Steven N. Evans
6. Percy Liang
7. Behnam Neyshabur
8. Sergey Levine

Citations in the corpus

• Representing smooth functions as compositions of near-identity functions with implications for deep network optimization (2018, background): “(This was shown for generic parameter values and tanh networks of arbitrary depth by Fefferman (1994), and improved by Albertini and Sontag (1992) for the special case of two-layer networks.)”
• Bayesian Backpropagation Over I-O Functions Rather Than Weights (1993, background): “(Such symmetries arise from things like permuting the hidden neurons or changing the sign of all weights leading into a hidden neuron -see (Fefferman, 1993) and references therein.)”