Efficient Generation of Structured Objects with Constrained Adversarial Networks

by   Luca Di Liello, et al.
Università di Trento

Generative Adversarial Networks (GANs) struggle to generate structured objects like molecules and game maps. The issue is that structured objects must satisfy hard requirements (e.g., molecules must be chemically valid) that are difficult to acquire from examples alone. As a remedy, we propose Constrained Adversarial Networks (CANs), an extension of GANs in which the constraints are embedded into the model during training. This is achieved by penalizing the generator proportionally to the mass it allocates to invalid structures. In contrast to other generative models, CANs support efficient inference of valid structures (with high probability) and allows to turn on and off the learned constraints at inference time. CANs handle arbitrary logical constraints and leverage knowledge compilation techniques to efficiently evaluate the disagreement between the model and the constraints. Our setup is further extended to hybrid logical-neural constraints for capturing very complex constraints, like graph reachability. An extensive empirical analysis shows that CANs efficiently generate valid structures that are both high-quality and novel.


Labeled Graph Generative Adversarial Networks

As a new way to train generative models, generative adversarial networks...

Utilizing Generative Adversarial Networks for Stable Structure Generation in Angry Birds

This paper investigates the suitability of using Generative Adversarial ...

Physics-Constrained Generative Adversarial Networks for 3D Turbulence

Generative Adversarial Networks (GANs) have received wide acclaim among ...

Enforcing Deterministic Constraints on Generative Adversarial Networks for Emulating Physical Systems

Generative adversarial networks (GANs) are initially proposed to generat...

Discrete Object Generation with Reversible Inductive Construction

The success of generative modeling in continuous domains has led to a su...

GANs for LIFE: Generative Adversarial Networks for Likelihood Free Inference

We introduce a framework using Generative Adversarial Networks (GANs) fo...

Algorithms that get old : the case of generative algorithms

Generative IA networks, like the Variational Auto-Encoders (VAE), and Ge...

Please sign up or login with your details

Forgot password? Click here to reset