Value-based Search in Execution Space for Mapping Instructions to Programs

@inproceedings{Muhlgay2018ValuebasedSI,
  title={Value-based Search in Execution Space for Mapping Instructions to Programs},
  author={Dor Muhlgay and Jonathan Herzig and Jonathan Berant},
  booktitle={NAACL-HLT},
  year={2018}
}
Highlight Information
Training models to map natural language instructions to programs given target world supervision only requires searching for good programs at training time. [...] Key Method We then score search states on the beam by interpolating their expected reward with the likelihood of programs represented by the search state. Moreover, we search not in the space of programs but in a more compressed state of program executions, augmented with recent entities and actions. On the SCONE dataset, we show that our algorithm…Expand Abstract
2
Twitter Mentions

Figures, Tables, and Topics from this paper.

Citations

Publications citing this paper.

References

Publications referenced by this paper.
SHOWING 1-10 OF 50 REFERENCES

An Actor-Critic Algorithm for Sequence Prediction

VIEW 4 EXCERPTS
HIGHLY INFLUENTIAL