DRS
A formal semantic representation framework derived from Discourse Representation Theory, designed to capture sentence- and discourse-level semantics, including quantifier scope, coreference relations, and temporal information. Its output is a nested logical graph structure that encodes the meaning of linguistic expressions across multiple levels of discourse.
Timeline (last 90 days)
This study proposes using structural automata to systematically address structural constraint problems in neural semantic parsing, validated across decoding, post-processing, and training stages of DRS parsing, while also evaluating decoding performance on the AMR task
Under strict length constraints, APD improves DRS task performance over local masking by 5.1 and 2.4 points on the Smatch++ metric
When sampling eight candidates without constraints from the model, 10.2% of inputs cannot produce any valid output under limited training data, and this proportion drops to 0.4% when using the full training data
DRS originates from Discourse Representation Theory and is a nested logical graph structure for capturing sentence- and discourse-level semantics, including quantifier scope, coreference relations, and tense information
All Facts (4)
This study proposes using structural automata to systematically address structural constraint problems in neural semantic parsing, validated across decoding, post-processing, and training stages of DRS parsing, while also evaluating decoding performance on the AMR task
50%UnverifiedUnder strict length constraints, APD improves DRS task performance over local masking by 5.1 and 2.4 points on the Smatch++ metric
50%UnverifiedWhen sampling eight candidates without constraints from the model, 10.2% of inputs cannot produce any valid output under limited training data, and this proportion drops to 0.4% when using the full training data
50%UnverifiedDRS originates from Discourse Representation Theory and is a nested logical graph structure for capturing sentence- and discourse-level semantics, including quantifier scope, coreference relations, and tense information
50%