Vera Demberg

4 paper cards in Mova's library · 2012–2024

Vera Demberg wrote 4 of the papers Mova reads, in 2012–2024, most of them on Discourse analysis, Implicature, Treebanks; often with Frances Yung, Merel Scholman, Miriam Kaeshammer.

The profile is the library's match of the name; two people with one name are not always told apart.

Papers

RSA-Control: A Pragmatics-Grounded Lightweight Controllable Text Generation Framework (2024)
RSA-Control steers a pretrained language model towards a target attribute (for example non-toxic, or easy to read) without training a classifier.
SIGA: A Naturalistic NLI Dataset of English Scalar Implicatures with Gradable Adjectives (2024)
SIGA is a set of naturally occurring scalar implicatures with gradable adjectives. The authors mine the C4 web corpus for the reinforcement frame «WEAK but not STRONG» («good but not great»), using 88 adjective scale pairs from de Melo & Bansal (2013). The frame guarantees which scale the writer had in mind. Deleting…
DiscoGeM: A Crowdsourced Corpus of Genre-Mixed Implicit Discourse Relations (2022)
DiscoGeM holds 6,505 implicit relations between consecutive sentences (no connective among the first five words of the second) from political speeches (Europarl), 20 novels and Wikipedia summaries, each labelled by 10 crowd workers (65,863 judgements) with PDTB 3.0 senses.
German and English Treebanks and Lexica for Tree-Adjoining Grammars (2012)
The paper builds a treebank and a lexicon for PLTAG, an incremental, psycholinguistically motivated version of tree-adjoining grammar.