brainsteam.co.uk/brainsteam/content/replies/2022/11/23/1669233264.md

date	hypothesis-meta	in-reply-to	tags	type	url
2022-11-23T19:54:24	created document flagged group hidden id links permissions tags target text updated uri user user_info 2022-11-23T19:54:24.332809+00:00 title 2210.07188.pdf false __world__ false oTGKsmtoEe2RF0-NK45jew html incontext json https://hypothes.is/a/oTGKsmtoEe2RF0-NK45jew https://hyp.is/oTGKsmtoEe2RF0-NK45jew/arxiv.org/pdf/2210.07188.pdf https://hypothes.is/api/annotations/oTGKsmtoEe2RF0-NK45jew admin delete read update acct:ravenscroftj@hypothes.is acct:ravenscroftj@hypothes.is group:__world__ acct:ravenscroftj@hypothes.is coreference NLProc data-annotation selector source end start type 5934 5221 TextPositionSelector exact prefix suffix type Specifically, our work investigates the quality ofcrowdsourced coreference annotations when anno-tators are taught only simple coreference cases thatare treated uniformly across existing datasets (e.g.,pronouns). By providing only these simple cases,we are able to teach the annotators the concept ofcoreference, while allowing them to freely interpretcases treated differently across the existing datasets.This setup allows us to identify cases where ourannotators disagree among each other, but moreimportantly cases where they unanimously agreewith each other but disagree with the expert, thussuggesting cases that should be revisited by theresearch community when curating future unifiedannotation guidelines ficient payment-based platforms. .Our main contributions are:1. W TextQuoteSelector https://arxiv.org/pdf/2210.07188.pdf The aim of the work is to examine a simplified subset of co-reference phenomena which are generally treated the same across different existing datasets. This makes spotting inter-annotator disagreement easier - presumably because for simpler cases there are fewer modes of failure? 2022-11-23T19:54:24.332809+00:00 https://arxiv.org/pdf/2210.07188.pdf acct:ravenscroftj@hypothes.is display_name James Ravenscroft	https://arxiv.org/pdf/2210.07188.pdf	coreference NLProc data-annotation hypothesis	reply	/replies/2022/11/23/1669233264

Specifically, our work investigates the quality ofcrowdsourced coreference annotations when anno-tators are taught only simple coreference cases thatare treated uniformly across existing datasets (e.g.,pronouns). By providing only these simple cases,we are able to teach the annotators the concept ofcoreference, while allowing them to freely interpretcases treated differently across the existing datasets.This setup allows us to identify cases where ourannotators disagree among each other, but moreimportantly cases where they unanimously agreewith each other but disagree with the expert, thussuggesting cases that should be revisited by theresearch community when curating future unifiedannotation guidelines

The aim of the work is to examine a simplified subset of co-reference phenomena which are generally treated the same across different existing datasets.

This makes spotting inter-annotator disagreement easier - presumably because for simpler cases there are fewer modes of failure?