Feature Request: Maximum A Posteriori (MAP) estimation

tensorflow / probability

Probabilistic reasoning and statistical analysis in TensorFlow

https://www.tensorflow.org/probability/

Apache License 2.0

4.26k stars 1.1k forks source link

Feature Request: Maximum A Posteriori (MAP) estimation #43

Open matthew-mcateer opened 6 years ago

matthew-mcateer commented 6 years ago

For mixed models, it would be great if MCMC could be initialized with a more accurate estimation. This was a feature in Pymc and Edwardlib, but does not appear to be available in either tensorflow_probability or the edward2.

rlouf commented 6 years ago

Is there any interest in this in tfp? I’m currently working on the Stochastick Block Model example and I ended up re-implementing edward’s MAP. That would be a nice by-product.