Offline Reinforcement Learning approaches like Inverse Reinforcement Learning/ Batch RL/ imitation learning/ behavior cloning allow us to use previous demonstrations by an expert to learn a policy.
Many of the papers that I have found use expert data generated from a simulated environment for model training. However, I have some real world images and corresponding action that the expert took in these states e.g. images from a camera mounted on the car and the steering angle as an action. The action space is continuous meaning how fast the steering angle was applied.
Can I use this dataset as an expert demonstration in place of simulated dataset for training an Inverse Reinforcement learning algorithm?
