Investigating Independence vs. Control: Agenda-Setting in Russian News Coverage on Social Media

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 534–542 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 534 Negation Detection in Dutch Spoken Human-Computer Conversations Tom Sweers, Iris Hendrickx and Helmer Strik Centre for Language Studies, Centre for Language and Speech Technology Radboud University, Erasmusplein 1, 6500 HD, Nijmegen, The Netherlands tom.sweers@hotmail.com, iris.hendrickx@ru.nl, helmer.strik@ru.nl Abstract Proper recognition and interpretation of negation signals in text or communication is crucial for any form of full natural language understanding. It is also essential for computational approaches to natural language processing. In this study we focus on negation detection in Dutch spoken human-computer conversations. Since there exists no Dutch (dialogue) corpus annotated for negation we have annotated a Dutch corpus sample to evaluate our method for automatic negation detection. We use transfer learning and trained NegBERT (an existing BERT implementation used for negation detection) on English data with multilingual BERT to detect negation in Dutch dialogues. Our results show that adding in-domain training material improves the results. We show that we can detect both negation cues and scope in Dutch dialogues with high precision and recall. We provide a detailed error analysis and discuss the effects of cross-lingual and cross-domain transfer learning on automatic negation detection. Keywords: negation detection, corpus annotation, transfer learning 1. Introduction Automatic negation detection is a crucial building block within the dialogue manager of a conversational agent as the one developed in the BLISS project1 (van Waterschoot et al., 2020). The dialogue manager needs to perform proper negation handling both from a com- municative perspective and for content extraction. For example, when a user while talking with an agent says ‘i dont understand what you mean’ as response to a question from the agent, this should be recognized as a misunderstanding and not as an answer to the ques- tion. Various studies have shown the effectiveness of auto- mated negation detection and scope resolution in En- glish in text mining applications (e.g. (Li and Lu, 2018; Gautam et al., 2018; Chen, 2019; Khandelwal and Sawant, 2020)). This paper presents the first study on the applicability of negation cue detection and scope resolution in Dutch spoken dialogues. Since there ex- ists no spoken Dutch labeled negation corpus, a small Dutch dialogue corpus was annotated with negation cues (negation expressions) and their scopes. This corpus sample is publicly available for research. To counter the relatively small size of the corpus, transfer learning was used by training models on English cor- pora from different domains. Therefore this study also investigates the applicability of cross-lingual and cross- domain transfer learning in negation cue detection and scope resolution. We discuss negation from an NLP perspective. In Sec- tion 2 we first discuss related work on models for au- tomatic negation detection and corpora annotated with negation in Section 3. Section 4 presents the annotation guidelines and the Dutch spoken corpus sample anno- tated with negation. We present the models, baselines 1http://bliss.ruhosting.nl and experimental setup in …

Content truncated to 500 words