Fetching the paper…

Reading the bibliography…

2023

Dialogue Shaping: Empowering Agents through NPC Interaction

Zhou, Wei, Peng, Xiangyu, Riedl, Mark

Original
PDF
Post on X

Understand

One major challenge in reinforcement learning (RL) is the large amount of steps for the RL agent needs to converge in the training process and learn the optimal policy, especially in text-based game environments where the action space is extensive.

However, non-player characters (NPCs) sometimes hold some key information about the game, which can potentially help to train RL agents faster.
Thus, this paper explores how to interact and converse with NPC agents to get the key information using large language models (LLMs), as well as incorporate this information to speed up RL agent's training using knowledge graphs (KGs) and Story Shaping.

Dialogue Shaping: Empowering Agents through NPC Interaction · Around

Built on

Asynchronous methods for deep reinforcement learning,
V. Mnih, A. P. Badia, M. Mirza, A. Graves, T. Lillicrap, T. Harley, D. Silver, K. Kavukcuoglu, · 1937
Earlier work this paper cites.
S. Yao, R. Rao, M. Hausknecht, K. Narasimhan, Keep calm and explore: Language models for action generation in text-based games, 2020. arXiv:2010.02903
Original
2010
Earlier work this paper cites.
Playing text-adventure games with graph-based deep reinforcement learning,
Original
P. Ammanabrolu, M. O. Riedl, · 2018
Earlier work this paper cites.
Learning to speak and act in a fantasy text adventure game,
Original
J. Urbanek, A. Fan, S. Karamcheti, S. Jain, S. Humeau, E. Dinan, T. Rocktäschel, D. Kiela, A. Szlam, J. Weston, · 2019
Earlier work this paper cites.

Similar

Bringing stories alive: Generating interactive fiction worlds,
P. Ammanabrolu, W. Cheung, D. Tu, W. Broniec, M. Riedl, · 2020
Cited alongside, same era.
Graph constrained reinforcement learning for natural language action spaces,
Original
P. Ammanabrolu, M. Hausknecht, · 2020
Cited alongside, same era.
How to avoid being eaten by a grue: Structured exploration strategies for textual worlds,
Original
P. Ammanabrolu, E. Tien, M. Hausknecht, M. O. Riedl, · 2020
Cited alongside, same era.
Deep reinforcement learning with stacked hierarchical attention for text-based games,
Y. Xu, M. Fang, L. Chen, Y. Du, J. T. Zhou, C. Zhang, · 2020
Cited alongside, same era.

Then

OpenAI, Chatgpt: A large-scale open-domain chatbot, https://openai.com/blog/chatgpt/ , 2022
2022
Later among the works it cites.
Inherently explainable reinforcement learning in natural language,
X. Peng, M. Riedl, P. Ammanabrolu, · 2022
Later among the works it cites.
Story shaping: Teaching agents human-like behavior with stories,
Original
X. Peng, C. Cui, W. Zhou, R. Jia, M. Riedl, · 2023
Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…