Memory-Augmented Neural Networks

This project contains implementations of memory augmented neural networks. This includes code in the following subdirectories:

• MemN2N-lang-model: This code trains MemN2N model for language modeling, see Section 5 of the paper "End-To-End Memory Networks". This code is implemented in Torch7 (written in Lua); more documentation is given in the README in that subdirectory.

• MemN2N-babi-matlab: The code for the MemN2N bAbI task experiments of Section 4 of the paper "End-To-End Memory Networks". This code is implemented in Matlab; more documentation is given in the README in that subdirectory.

• DBLL: Code to train MemN2N on tasks from the paper "Dialog-based Language Learning". This code is implemented in Torch7; more documentation is given in the README in that subdirectory.

• HITL: Code to train MemN2N on tasks from the paper "Dialogue Learning With Human-in-the-Loop". This code is implemented in Torch7; more documentation is given in the README in that subdirectory.

• AskingQuestions: Code to train MemN2N on tasks from the paper "Learning through Dialogue Interactions". This code is implemented in Torch7; more documentation is given in the README in that subdirectory.

• KVmemnn: Code to train MemN2N on tasks from the paper "Key-Value Memory Networks for Directly Reading Documents". This code is implemented in Torch7; more documentation is given in the README in that subdirectory.

• EntNet-babi: Code to train an Entity Network on bAbI tasks, as described in the paper "Tracking the World State with Recurrent Entity Networks". This code is implemented in Torch7; more documentation is given in the README in that subdirectory.

Other 3rd party implementations

• python-babi: MemN2N implementation on bAbI tasks with very nice interactive demo.
• theano-babi: MemN2N implementation in Theano for bAbI tasks.
• tf-lang: MemN2N language model implementation in TensorFlow.
• tf-babi: Another MemN2N implementation of MemN2N in TensorFlow, but for bAbI tasks.