Tuan Anh Le

analysis-by-synthesis by learning to invert generative black boxes, wip

19 April 2017

notes on (Nair et al., 2008).

understanding: 5/10
code: ?



goal: approximate posterior \(q_{\phi}(x \given y) \approx p(x \given y)\). (not exactly posterior since they have no defined prior).


  1. pick \(y_n\) from training data set.
  2. run it through the recognition and perturb to obtain \(x\).
  3. obtain \(y\) from black box generator.
  4. perform supervised learning on \((x, y)\):
    1. get \(x'\) from recognition network.
    2. calculate loss between \(x\) and \(x'\).
    3. take gradient step.


  1. Nair, V., Susskind, J., & Hinton, G. E. (2008). Analysis-bySynthesis by Learning to Invert Generative Black Boxes.
      title = {Analysis-bySynthesis by Learning to Invert Generative Black Boxes},
      author = {Nair, Vinod and Susskind, Josh and Hinton, Geoffrey E.},
      year = {2008}