Discontinuous Constituent Parsing as Sequence Labeling
This addresses a specific problem in natural language processing for researchers and practitioners, representing an incremental advancement in parsing techniques.
The paper tackled the problem of discontinuous constituent parsing by reducing it to sequence labeling, showing that existing reductions don't support discontinuities and proposing to encode tree discontinuities as nearly ordered permutations. The experiments demonstrated that with the right representation, the models achieved fast and accurate performance.
This paper reduces discontinuous parsing to sequence labeling. It first shows that existing reductions for constituent parsing as labeling do not support discontinuities. Second, it fills this gap and proposes to encode tree discontinuities as nearly ordered permutations of the input sequence. Third, it studies whether such discontinuous representations are learnable. The experiments show that despite the architectural simplicity, under the right representation, the models are fast and accurate.