Elaborating on Learned Demonstrations with Temporal Logic Specifications

@article{Innes2020ElaboratingOL,
  title={Elaborating on Learned Demonstrations with Temporal Logic Specifications},
  author={Craig Innes and Subramanian Ramamoorthy},
  journal={ArXiv},
  year={2020},
  volume={abs/2002.00784}
}
  • Craig Innes, Subramanian Ramamoorthy
  • Published 2020
  • Mathematics, Computer Science
  • ArXiv
  • Most current methods for learning from demonstrations assume that those demonstrations alone are sufficient to learn the underlying task. This is often untrue, especially if extra safety specifications exist which were not present in the original demonstrations. In this paper, we allow an expert to elaborate on their original demonstration with additional specification information using linear temporal logic (LTL). Our system converts LTL specifications into a differentiable loss. This loss is… CONTINUE READING

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 32 REFERENCES

    DL2: Training and Querying Neural Networks with Logic

    VIEW 4 EXCERPTS
    HIGHLY INFLUENTIAL

    A Policy Search Method For Temporal Logic Specified Reinforcement Learning Tasks

    VIEW 1 EXCERPT

    Adam: A Method for Stochastic Optimization

    VIEW 1 EXCERPT

    Correct, Reactive, High-Level Robot Control

    VIEW 2 EXCERPTS

    Deep Encoder-Decoder Networks for Mapping Raw Images to Dynamic Movement Primitives

    VIEW 2 EXCERPTS