WilliamKulp / alta2017-rl

Code related to the ALTA 2017 paper "Towards the Use of Deep Reinforcement Learning with Global Policy For Query-based Extractive Summarisation"

bioasq-rl

Experiments using Reinforcement Learning for BioASQ as described in the following paper:

D. Molla (2017). Towards the use of deep reinforce learning with global policy for query-based extractive summarisation. Proceedings ALTA 2017, Brisbane, Australia.

Train and evaluate reinforce for debugging

$ python3
>>> import rl
>>> rl.DEBUG = True
>>> from rl import reinforce
>>> reinforce.train()

Train and evaluate reinforce for production

>>> from rl import reinforce
>>> reinforce.train()

Run "first n" baseline

$ python3
>>> from rl import reinforce
>>> reinforce.baseline()