Learning Retrospective Knowledge with Reverse Reinforcement Learning
This work addresses a gap in reinforcement learning for representing past-oriented knowledge, which could benefit domains like robotics or diagnostics, but it appears incremental as it extends existing GVF methods.
The authors tackled the problem of representing retrospective knowledge, which involves answering questions about past events' influence on the present, by introducing Reverse General Value Functions (GVFs) trained via Reverse Reinforcement Learning. They demonstrated the utility of Reverse GVFs in representation learning and anomaly detection, though no concrete numbers were provided in the abstract.
We present a Reverse Reinforcement Learning (Reverse RL) approach for representing retrospective knowledge. General Value Functions (GVFs) have enjoyed great success in representing predictive knowledge, i.e., answering questions about possible future outcomes such as "how much fuel will be consumed in expectation if we drive from A to B?". GVFs, however, cannot answer questions like "how much fuel do we expect a car to have given it is at B at time $t$?". To answer this question, we need to know when that car had a full tank and how that car came to B. Since such questions emphasize the influence of possible past events on the present, we refer to their answers as retrospective knowledge. In this paper, we show how to represent retrospective knowledge with Reverse GVFs, which are trained via Reverse RL. We demonstrate empirically the utility of Reverse GVFs in both representation learning and anomaly detection.