Exploiting Language Instructions for Interpretable and Compositional Reinforcement Learning
van der Meer, Michiel, Pirotta, Matteo, Bruni, Elia
–arXiv.org Artificial Intelligence
In this work, we present an alternative approach to making an agent compositional through the use of a diagnostic classifier. Because of the need for explainable agents in automated decision processes, we attempt to interpret the latent space from an RL agent to identify its current objective in a complex language instruction. Results show that the classification process causes changes in the hidden states which makes them more easily interpretable, but also causes a shift in zero-shot performance to novel instructions. Lastly, we limit the supervisory signal on the classification, and observe a similar but less notable effect.
arXiv.org Artificial Intelligence
Jan-13-2020
- Country:
- North America > United States
- Massachusetts (0.04)
- Europe
- Sweden > Stockholm
- Stockholm (0.04)
- Netherlands > North Holland
- Amsterdam (0.04)
- Sweden > Stockholm
- North America > United States
- Genre:
- Research Report > New Finding (0.48)
- Industry:
- Education (0.68)
- Leisure & Entertainment > Games
- Computer Games (0.46)
- Technology: