Combating False Negatives in Adversarial Imitation Learning
This addresses a specific bottleneck in imitation learning for AI agents, offering incremental improvements in sample efficiency.
The paper tackles the problem of false negatives in adversarial imitation learning, where successful agent episodes are incorrectly labeled as negative by the discriminator, hindering learning; it proposes a method that improves sample efficiency by at least an order of magnitude in the BabyAI environment.
In adversarial imitation learning, a discriminator is trained to differentiate agent episodes from expert demonstrations representing the desired behavior. However, as the trained policy learns to be more successful, the negative examples (the ones produced by the agent) become increasingly similar to expert ones. Despite the fact that the task is successfully accomplished in some of the agent's trajectories, the discriminator is trained to output low values for them. We hypothesize that this inconsistent training signal for the discriminator can impede its learning, and consequently leads to worse overall performance of the agent. We show experimental evidence for this hypothesis and that the 'False Negatives' (i.e. successful agent episodes) significantly hinder adversarial imitation learning, which is the first contribution of this paper. Then, we propose a method to alleviate the impact of false negatives and test it on the BabyAI environment. This method consistently improves sample efficiency over the baselines by at least an order of magnitude.