Teach agents to operate in complex text-based simulations
Two key components in creating effective language-based AI agents are interactivity and environment grounding, shown to be vital parts of language learning in humans, and posit that interactive narratives should be the environments of choice for training such agents. These games are simulations in which an agent interacts with the world through natural language—”perceiving’’, “acting upon’’, and “talking to’’ the world using textual descriptions, commands, and dialogue—and as such exist at the intersection of natural language processing, storytelling, and sequential decision making.
References
2021
Situated Language Learning via Interactive Narratives
@article{ammanabrolu2021situated,title={Situated Language Learning via Interactive Narratives},author={Ammanabrolu, Prithviraj and Riedl, Mark O},journal={Patterns, Cell Press},volume={},url={https://www.cell.com/patterns/fulltext/S2666-3899(21)00159-8},year={2021},}
2020
Interactive fiction games: A colossal adventure
Matthew Hausknecht, Prithviraj Ammanabrolu, Marc-Alexandre Côté, and Xingdi Yuan
In Proceedings of the AAAI Conference on Artificial Intelligence, 2020
@inproceedings{hausknecht2020interactive,title={Interactive fiction games: A colossal adventure},author={Hausknecht, Matthew and Ammanabrolu, Prithviraj and C{\^o}t{\'e}, Marc-Alexandre and Yuan, Xingdi},booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},volume={34},number={05},pages={7903--7910},year={2020},url={https://arxiv.org/abs/1909.05398},}
Graph Constrained Reinforcement Learning for Natural Language Action Spaces
Prithviraj Ammanabrolu, and Matthew Hausknecht
In International Conference on Learning Representations, 2020
@inproceedings{ammanabrolu2020graph,title={Graph Constrained Reinforcement Learning for Natural Language Action Spaces},author={Ammanabrolu, Prithviraj and Hausknecht, Matthew},booktitle={International Conference on Learning Representations},year={2020},url={https://openreview.net/forum?id=B1x6w0EtwH},}
How to avoid being eaten by a grue: Structured exploration strategies for textual worlds
Prithviraj Ammanabrolu, Ethan Tien, Matthew Hausknecht, and Mark O Riedl
@article{ammanabrolu2020avoid,title={How to avoid being eaten by a grue: Structured exploration strategies for textual worlds},author={Ammanabrolu, Prithviraj and Tien, Ethan and Hausknecht, Matthew and Riedl, Mark O},journal={arXiv preprint arXiv:2006.07409},url={https://arxiv.org/abs/2006.07409},year={2020},}
2019
Playing Text-Adventure Games with Graph-Based Deep Reinforcement Learning
Prithviraj Ammanabrolu, and Mark Riedl
In North American Chapter of the Association for Computational Linguistics (NAACL-HLT) 2019, 2019
@inproceedings{ammanabrolu2019playing,title={Playing Text-Adventure Games with Graph-Based Deep Reinforcement Learning},author={Ammanabrolu, Prithviraj and Riedl, Mark},booktitle={North American Chapter of the Association for Computational Linguistics (NAACL-HLT) 2019},url={https://aclanthology.org/N19-1358/},year={2019},}