Repository: writer stance - 1000 German texts from the X-stance corpus (see References X-stance) annotated for writer stance - two formats: conll and BIO (sort of) - see below - goal: explicit and implicit writer stance wrt. to entities (nouns) and (some) events (verbs) - implicit and explicit are not distinguished (no indication which case was annotated) - labels: BIO Conll - in favour: PRO p - against: CON c - neutral: O no label - 3 Annotations (see References DeInStance) Formats: conll: (parzu parser) 1 Das die ART ART Def|Neut|Nom|Sg 2 det _ _ p2 Arbeitsgesetz Arbeitsgesetz N NN Neut|Nom|Sg 3 subj _ _ 3 regelt regeln V VVFIN 3|Sg|Pres|Ind 0 root _ _ 4 die die ART ART Def|Fem|Acc|Pl 5 det _ _ 5 Arbeitszeiten Arbeitszeit N NN Fem|Acc|Pl 3 obja _ _ 6 und und KON KON _ 3 kon _ _ 7 schützt schützen V VVFIN _|_|Pres|Ind 6 cj _ _ 8 den die ART ART Def|Masc|Acc|Sg 9 det _ _ p9 Arbeitnehmer Arbeitnehmer N NN Masc|Acc|Sg 7 obja _ _ 10 . . $. $. _ 0 root _ _ head was annotated directly in front of the index (1. column) e.g. "p2 Arbeitsgesetz Arbeitsgesetz N NN Neut|Nom|Sg 3 subj _ _" i.e. the writer is in favour of "Arbeitsgesetz" BIO: wordform lemma label Das die O Arbeitsgesetz Arbeitsgesetz PRO regelt regeln O die die O Arbeitszeiten Arbeitszeit O und und O schützt schützen O den die O Arbeitnehmer Arbeitnehmer PRO . . O head was annotated (last column) e.g. "Arbeitsgesetz Arbeitsgesetz PRO" for X-stance corpus see also https://vamvas.ch/more-general-stance-detection-with-x-stance References: @article{X-stance, author = {Jannis Vamvas and Rico Sennrich}, title = {X-Stance: {A} Multilingual Multi-Target Dataset for Stance Detection}, journal = {CoRR}, volume = {abs/2003.08385}, year = {2020}, url = {https://arxiv.org/abs/2003.08385}, eprinttype = {arXiv}, eprint = {2003.08385}, timestamp = {Tue, 24 Mar 2020 16:42:29 +0100}, biburl = {https://dblp.org/rec/journals/corr/abs-2003-08385.bib}, bibsource = {dblp computer science bibliography, https://dblp.org} } @inproceedings{DeInStance, booktitle = {17th Conference on Natural Language Processing (KONVENS)}, month = {September}, title = {DeInStance: Creating and Evaluating a {G}erman Corpus for Fine-Grained Inferred Stance Detection}, author = {Anne Gohring and Manfred Klenner and Sophia Conrad}, publisher = {ACL Anthology}, year = {2021}, pages = {213--217}, language = {english}, url = {https://doi.org/10.5167/uzh-207940}, abstract = {We introduce deInStance, a corpus of 1000 politicians? answers in German (de) containing sentences labeled with explicitly expressed and inferred stances - pro and con relations - by 3 annotators. They achieved an acceptable inter-rater agreement given the inherent subjective nature of the task. A first baseline, a fine-tuned BERT-based token classifier, achieved F1-scores of around 70\% . Our focus is on the difficult subclass of sentences comprising only non-polar words, but still with an (implicit) pro or con perspective of the writer.} }