Structured Mean-Field Variational Inference for Higher-Order Span-Based Semantic Role

In this work, we enhance higher-order graphbased approaches for span-based semantic role labeling (SRL) by means of structured modeling. To decrease the complexity of higher-order modeling, we decompose the edge from predicate word to argument span into three different edges, predicate-to-head (P2H), predicateto-tail (P2T), and head-to-tail (H2T), where head/tail means the first/last word of the semantic argument span. As such, we use a CRF-based higher-order dependency parser and leverage Mean-Field Variational Inference (MFVI) for higher-order inference. Moreover, since semantic arguments of predicates are often constituents within a constituency parse tree, we can leverage such nice structural property by defining a TreeCRF distribution over all H2T edges, using the idea of partial marginalization to define structural training loss. We further leverage structured MFVI to enhance inference. We experiment on spanbased SRL benchmarks, showing the effectiveness of both higher-order and structured modeling and the combination thereof. In addition, we show superior performance of structured MFVI against vanilla MFVI. Our code is publicly available at https://github.com/ VPeterV/Structured-MFVI .