Meta-Optimizing Semantic Evolutionary Search

1 code implementation • Association for Computing Machinery (ACM) 2017 • Moshe Looks

I present MOSES (meta-optimizing semantic evolutionary search), a new probabilistic modeling (estimation of distribution) approach to program evolution.

Problem Decomposition

Deep Learning with Dynamic Computation Graphs

2 code implementations • 7 Feb 2017 • Moshe Looks, Marcello Herreshoff, DeLesley Hutchins, Peter Norvig

However, since the computation graph has a different shape and size for every input, such networks do not directly support batched training or inference.

A Minimalistic Approach to Sum-Product Network Learning for Real Applications

no code implementations • 12 Feb 2016 • Viktoriya Krakovna, Moshe Looks

Sum-Product Networks (SPNs) are a class of expressive yet tractable hierarchical graphical models.

Clustering