Semantic Roles and Frame Semantics

Semantic Roles and Frame Semantics

Numbered arguments and frames

Labeling roles from a shallow parse

What arguments are plausible

Using the frameset where ARG0 is the loader, ARG1 the cargo, ARG2 the destination and ARG3 the source, what roles do the crew, the crates and the pier carry in the crew unloaded the crates at the pier?

In the sails were stowed by the skipper, which mapping of surface positions to the frameset roles is correct if ARG0 is the stower and ARG1 the thing stowed?

A semantic role labeler is evaluated on gold spans and labels, achieving labeled precision $0.868$ and labeled recall $0.892$. Compute its labeled F1, rounding to three decimals.

The change-of-position-on-a-scale frame collects rise, fall, increase and their nominal forms. What does this buy over per-verb PropBank-style framesets?

A local role classifier labels two constituents of one predicate both as ARG0, and one of them overlaps the span of the other. Why is this a global problem, and what fixes it?