Learning and Evaluating Emotion Lexicons for 91 Languages

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 1202–1217 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 1202 Learning and Evaluating Emotion Lexicons for 91 Languages Sven Buechel, Susanna R¨ucker, and Udo Hahn {sven.buechel|susanna.ruecker|udo.hahn}@uni-jena.de Jena University Language and Information Engineering (JULIE) Lab Friedrich-Schiller-Universit¨at Jena, Jena, Germany https://julielab.de Abstract Emotion lexicons describe the affective mean- ing of words and thus constitute a centerpiece for advanced sentiment and emotion analysis. Yet, manually curated lexicons are only avail- able for a handful of languages, leaving most languages of the world without such a precious resource for downstream applications. Even worse, their coverage is often limited both in terms of the lexical units they contain and the emotional variables they feature. In or- der to break this bottleneck, we here intro- duce a methodology for creating almost arbi- trarily large emotion lexicons for any target language. Our approach requires nothing but a source language emotion lexicon, a bilin- gual word translation model, and a target lan- guage embedding model. Fulfilling these re- quirements for 91 languages, we are able to generate representationally rich high-coverage lexicons comprising eight emotional variables with more than 100k lexical entries each. We evaluated the automatically generated lexicons against human judgment from 26 datasets, spanning 12 typologically diverse languages, and found that our approach produces results in line with state-of-the-art monolingual ap- proaches to lexicon creation and even sur- passes human reliability for some languages and variables. Code and data are available at github.com/JULIELab/MEmoLon archived under DOI 10.5281/zenodo.3779901. 1 Introduction An emotion lexicon is a lexical repository which encodes the affective meaning of individual words (lexical entries). Most simply, affective meaning can be encoded in terms of polarity, i.e., the dis- tinction whether an item is considered as positive, negative, or neutral. This is the case for many well-known resources such as WORDNET-AFFECT (Strapparava and Valitutti, 2004), SENTIWORD- NET (Baccianella et al., 2010), or VADER (Hutto and Gilbert, 2014). Yet, an increasing number of researchers focus on more expressive encodings for affective states inspired by distinct lines of work in psychology (Yu et al., 2016; Buechel and Hahn, 2017; Sedoc et al., 2017; Abdul-Mageed and Un- gar, 2017; Bostan and Klinger, 2018; Mohammad, 2018; Troiano et al., 2019). Psychologists, on the one hand, value such lex- icons as a controlled set of stimuli for designing experiments, e.g., to investigate patterns of lexi- cal access or the structure of memory (Hofmann et al., 2009; Monnier and Syssau, 2008). NLP researchers, on the other hand, use them to aug- ment the emotional loading of word embeddings (Yu et al., 2017; Khosla et al., 2018), as addi- tional input to sentence-level emotion models so that the performance of even the most sophisti- cated neural network gets boosted (Mohammad and Bravo-Marquez, 2017; Mohammad et al., 2018; De Bruyne et al., 2019), or rely on them in a keyword-spotting approach when no training data is available, e.g., for studies dealing with historical language stages (Buechel et al., 2016). …

Content truncated to 500 words