Learning Algorithm & Program Induction

• Dynamic Neural Turing Machine with Soft and Hard Addressing Schemes [arxiv]
• Learning Neural Programs To Parse Programs [arxiv]
• Neural Symbolic Machines: Learning Semantic Parsers on Freebase with Weak Supervision [arxiv]

RL

• FeUdal Networks for Hierarchical Reinforcement Learning [arxiv]
• Feature Control as Intrinsic Motivation for Hierarchical Reinforcement Learning [arxiv]
• DARLA: Improving Zero-Shot Transfer in Reinforcement Learning [paper]

Reasoning & Learn visual concepts

• beta-VAE: Learning Basic Visual Concepts with a Constrained Variational Framework [openreview]
• Learning to Reason: End-to-End Module Networks for Visual Question Answering [arxiv]

GAN

• Good Semi-supervised Learning that Requires a Bad GAN [arxiv]