ml-ai-nlp-ir / n2n_dialog_system

Goal oriented dialog systems for bAbI tasks!

End-to-End Neural Networks for Goal Oriented Dialog

The objective of this project is to explore neural network architectures for question answering and goal oriented dialog systems. We will train on (6) dialog bAbI tasks. Tensorflow is used for building the models. I have listed some of the influential/interesting papers below. If I have missed something, let me know.

The (6) dialog bAbI tasks

This section presents the set of 6 tasks for testing end-to-end dialog systems in the restaurant domain. Each task tests a unique aspect of dialog.

Setup

# python2 is not supported
sudo -H pip3 install -r requirements.txt
# if this doesn't work, raise an issue

Learning End-to-End Goal-Oriented Dialog

# run main.py without arguments, for usage information
#  usage: main.py [-h] (-i | -t) [--task_id TASK_ID] [--batch_size BATCH_SIZE]
#               [--epochs EPOCHS] [--eval_interval EVAL_INTERVAL]
#               [--log_file LOG_FILE]
#  main.py: error: one of the arguments -i/--infer -t/--train is required
python3 main.py --train --task_id=3 --log_file=log.task3.txt

Results

Task	Training Accuracy	Validation Accuracy
1	100	99.7
2	100	100
3	100	74.71
4	100	56.67
5	100	98.42
6	76.61	47.08

Papers

1. Learning End-to-End Goal-Oriented Dialog, review
2. A Copy-Augmented Sequence-to-Sequence Architecture Gives Good Performance on Task-Oriented Dialogue
3. Hybrid Code Networks: practical and efficient end-to-end dialog control with supervised and reinforcement learning
4. Gated End-to-End Memory Networks
5. Query-Reduction Networks for Question Answering
6. Ask Me Even More: Dynamic Memory Tensor Networks

Jargons

• OOV : Out Of Vocabulary
• hops : number of memory lookups

Credits

• Tweaked vyraun's chatbot-MemN2N-tensorflow repository, for using Memory Networks for bAbI dialog tasks