The success of generative modeling in continuous domains has led to a surge
of interest in generating discrete data such as molecules, source code, and
graphs. However, construction histories for these discrete objects are
typically not unique and so generative models must reason about intractably
large spaces in order to learn. Additionally, structured discrete domains are
often characterized by strict constraints on what constitutes a valid object
and generative models must respect these requirements in order to produce
useful novel samples. Here, we present a generative model for discrete objects
employing a Markov chain where transitions are restricted to a set of local
operations that preserve validity. Building off of generative interpretations
of denoising autoencoders, the Markov chain alternates between producing 1) a
sequence of corrupted objects that are valid but not from the data
distribution, and 2) a learned reconstruction distribution that attempts to fix
the corruptions while also preserving validity. This approach constrains the
generative model to only produce valid objects, requires the learner to only
discover local modifications to the objects, and avoids marginalization over an
unknown and potentially large space of construction histories. We evaluate the
proposed approach on two highly structured discrete domains, molecules and
Laman graphs, and find that it compares favorably to alternative methods at
capturing distributional statistics for a host of semantically relevant
metrics.
Description
[1907.08268] Discrete Object Generation with Reversible Inductive Construction
%0 Conference Paper
%1 seff2019discrete
%A Seff, Ari
%A Zhou, Wenda
%A Damani, Farhan
%A Doyle, Abigail
%A Adams, Ryan P.
%D 2019
%K generative-models neurips2019 readings
%T Discrete Object Generation with Reversible Inductive Construction
%U http://arxiv.org/abs/1907.08268
%X The success of generative modeling in continuous domains has led to a surge
of interest in generating discrete data such as molecules, source code, and
graphs. However, construction histories for these discrete objects are
typically not unique and so generative models must reason about intractably
large spaces in order to learn. Additionally, structured discrete domains are
often characterized by strict constraints on what constitutes a valid object
and generative models must respect these requirements in order to produce
useful novel samples. Here, we present a generative model for discrete objects
employing a Markov chain where transitions are restricted to a set of local
operations that preserve validity. Building off of generative interpretations
of denoising autoencoders, the Markov chain alternates between producing 1) a
sequence of corrupted objects that are valid but not from the data
distribution, and 2) a learned reconstruction distribution that attempts to fix
the corruptions while also preserving validity. This approach constrains the
generative model to only produce valid objects, requires the learner to only
discover local modifications to the objects, and avoids marginalization over an
unknown and potentially large space of construction histories. We evaluate the
proposed approach on two highly structured discrete domains, molecules and
Laman graphs, and find that it compares favorably to alternative methods at
capturing distributional statistics for a host of semantically relevant
metrics.
@inproceedings{seff2019discrete,
abstract = {The success of generative modeling in continuous domains has led to a surge
of interest in generating discrete data such as molecules, source code, and
graphs. However, construction histories for these discrete objects are
typically not unique and so generative models must reason about intractably
large spaces in order to learn. Additionally, structured discrete domains are
often characterized by strict constraints on what constitutes a valid object
and generative models must respect these requirements in order to produce
useful novel samples. Here, we present a generative model for discrete objects
employing a Markov chain where transitions are restricted to a set of local
operations that preserve validity. Building off of generative interpretations
of denoising autoencoders, the Markov chain alternates between producing 1) a
sequence of corrupted objects that are valid but not from the data
distribution, and 2) a learned reconstruction distribution that attempts to fix
the corruptions while also preserving validity. This approach constrains the
generative model to only produce valid objects, requires the learner to only
discover local modifications to the objects, and avoids marginalization over an
unknown and potentially large space of construction histories. We evaluate the
proposed approach on two highly structured discrete domains, molecules and
Laman graphs, and find that it compares favorably to alternative methods at
capturing distributional statistics for a host of semantically relevant
metrics.},
added-at = {2019-12-11T00:34:47.000+0100},
author = {Seff, Ari and Zhou, Wenda and Damani, Farhan and Doyle, Abigail and Adams, Ryan P.},
biburl = {https://www.bibsonomy.org/bibtex/2ba9b19bc73c8169948f83c877f32bf21/kirk86},
description = {[1907.08268] Discrete Object Generation with Reversible Inductive Construction},
interhash = {ce857d73265a3a39bb1646d33b9ac76d},
intrahash = {ba9b19bc73c8169948f83c877f32bf21},
keywords = {generative-models neurips2019 readings},
note = {cite arxiv:1907.08268},
timestamp = {2019-12-11T00:34:47.000+0100},
title = {Discrete Object Generation with Reversible Inductive Construction},
url = {http://arxiv.org/abs/1907.08268},
year = 2019
}