Publications

Implicit discourse relation recognition is an extremely challenging task due to the lack of indicative connectives. Various neural network architectures have been proposed for this task recently, but most of them suffer from the shortage of labeled data. In this paper, we address this problem by procuring additional training data from parallel corpora: When humans translate a text, they sometimes add connectives (a process known as explicitation). We automatically back-translate it into an English connective and use it to infer a label with high confidence. We show that a training set several times larger than the original training set can be generated this way. With the extra labeled instances, we show that even a simple bidirectional Long Short-Term Memory Network can outperform the current state-of-the-art.

• https://aclanthology.org/I17-1049

@inproceedings{Shi2017b,
title = {Using explicit discourse connectives in translation for implicit discourse relation classification},
author = {Wei Shi and Frances Pik Yu Yung and Raphael Rubino and Vera Demberg},
url = {https://aclanthology.org/I17-1049},
year = {2017},
date = {2017},
booktitle = {Proceedings of the Eighth International Joint Conference on Natural Language Processing (Volume 1: Long Papers)},
pages = {484-495},
publisher = {Asian Federation of Natural Language Processing},
address = {Taipei, Taiwan},
abstract = {Implicit discourse relation recognition is an extremely challenging task due to the lack of indicative connectives. Various neural network architectures have been proposed for this task recently, but most of them suffer from the shortage of labeled data. In this paper, we address this problem by procuring additional training data from parallel corpora: When humans translate a text, they sometimes add connectives (a process known as explicitation). We automatically back-translate it into an English connective and use it to infer a label with high confidence. We show that a training set several times larger than the original training set can be generated this way. With the extra labeled instances, we show that even a simple bidirectional Long Short-Term Memory Network can outperform the current state-of-the-art.},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

The task of implicit discourse relation classification has received increased attention in recent years, including two CoNNL shared tasks on the topic. Existing machine learning models for the task train on sections 2-21 of the PDTB and test on section 23, which includes a total of 761 implicit discourse relations. In this paper, we{‚}d like to make a methodological point, arguing that the standard test set is too small to draw conclusions about whether the inclusion of certain features constitute a genuine improvement, or whether one got lucky with some properties of the test set, and argue for the adoption of cross validation for the discourse relation classification task by the community.

• https://aclanthology.org/E17-2024

@inproceedings{Shi2017,
title = {On the need of cross validation for discourse relation classification},
author = {Wei Shi and Vera Demberg},
url = {https://aclanthology.org/E17-2024},
year = {2017},
date = {2017},
booktitle = {Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 2, Short Papers},
pages = {150-156},
publisher = {Association for Computational Linguistics},
address = {Valencia, Spain},
abstract = {The task of implicit discourse relation classification has received increased attention in recent years, including two CoNNL shared tasks on the topic. Existing machine learning models for the task train on sections 2-21 of the PDTB and test on section 23, which includes a total of 761 implicit discourse relations. In this paper, we{'}d like to make a methodological point, arguing that the standard test set is too small to draw conclusions about whether the inclusion of certain features constitute a genuine improvement, or whether one got lucky with some properties of the test set, and argue for the adoption of cross validation for the discourse relation classification task by the community.},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

• https://www.sciencedirect.com/science/article/pii/S0749596X17300566

@article{Merel2017,
title = {"On the one hand" as a cue to anticipate upcoming discourse structure},
author = {Merel Scholman and Hannah Rohde and Vera Demberg},
url = {https://www.sciencedirect.com/science/article/pii/S0749596X17300566},
year = {2017},
date = {2017},
journal = {Journal of Memory and Language},
pages = {47-60},
volume = {97},
abstract = {

Copy BibTeX to Clipboard

• https://www.researchgate.net/publication/318569668_Examples_and_Specifications_that_Prove_a_Point_Identifying_Elaborative_and_Argumentative_Discourse_Relations

@article{Scholman2017,
title = {Examples and specifications that prove a point: Distinguishing between elaborative and argumentative discourse relations},
author = {Merel Scholman and Vera Demberg},
url = {https://www.researchgate.net/publication/318569668_Examples_and_Specifications_that_Prove_a_Point_Identifying_Elaborative_and_Argumentative_Discourse_Relations},
year = {2017},
date = {2017},
journal = {Dialogue and Discourse},
pages = {53-86},
volume = {8},
number = {2},
abstract = {

Copy BibTeX to Clipboard

Traditional discourse annotation tasks are considered costly and time-consuming, and the reliability and validity of these tasks is in question. In this paper, we investigate whether crowdsourcing can be used to obtain reliable discourse relation annotations. We also examine the influence of context on the reliability of the data. The results of a crowdsourced connective insertion task showed that the method can be used to obtain reliable annotations: The majority of the inserted connectives converged with the original label. Further, the method is sensitive to the fact that multiple senses can often be inferred for a single relation. Regarding the presence of context, the results show no significant difference in distributions of insertions between conditions overall. However, a by-item comparison revealed several characteristics of segments that determine whether the presence of context makes a difference in annotations. The findings discussed in this paper can be taken as evidence that crowdsourcing can be used as a valuable method to obtain insights into the sense(s) of relations.

• https://aclanthology.org/W17-0803
• https://doi.org/10.18653/v1/W17-0803

@inproceedings{Scholman2017,
title = {Crowdsourcing discourse interpretations: On the influence of context and the reliability of a connective insertion task},
author = {Merel Scholman and Vera Demberg},
url = {https://aclanthology.org/W17-0803},
doi = {https://doi.org/10.18653/v1/W17-0803},
year = {2017},
date = {2017},
booktitle = {Proceedings of the 11th Linguistic Annotation Workshop},
pages = {24-33},
publisher = {Association for Computational Linguistics},
address = {Valencia, Spain},
abstract = {Traditional discourse annotation tasks are considered costly and time-consuming, and the reliability and validity of these tasks is in question. In this paper, we investigate whether crowdsourcing can be used to obtain reliable discourse relation annotations. We also examine the influence of context on the reliability of the data. The results of a crowdsourced connective insertion task showed that the method can be used to obtain reliable annotations: The majority of the inserted connectives converged with the original label. Further, the method is sensitive to the fact that multiple senses can often be inferred for a single relation. Regarding the presence of context, the results show no significant difference in distributions of insertions between conditions overall. However, a by-item comparison revealed several characteristics of segments that determine whether the presence of context makes a difference in annotations. The findings discussed in this paper can be taken as evidence that crowdsourcing can be used as a valuable method to obtain insights into the sense(s) of relations.},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

Inferring implicit discourse relations in natural language text is the most difficult subtask in discourse parsing. Many neural network models have been proposed to tackle this problem. However, the comparison for this task is not unified, so we could hardly draw clear conclusions about the effectiveness of various architectures. Here, we propose neural network models that are based on feedforward and long-short term memory architecture and systematically study the effects of varying structures. To our surprise, the best-configured feedforward architecture outperforms LSTM-based model in most cases despite thorough tuning. Further, we compare our best feedforward system with competitive convolutional and recurrent networks and find that feedforward can actually be more effective. For the first time for this task, we compile and publish outputs from previous neural and non-neural systems to establish the standard for further comparison.

• https://aclanthology.org/E17-1027

@inproceedings{Rutherford2017,
title = {A systematic study of neural discourse models for implicit discourse relation},
author = {Attapol Rutherford and Vera Demberg and Nianwen Xue},
url = {https://aclanthology.org/E17-1027},
year = {2017},
date = {2017},
booktitle = {Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers},
pages = {281-291},
publisher = {Association for Computational Linguistics},
address = {Valencia, Spain},
abstract = {Inferring implicit discourse relations in natural language text is the most difficult subtask in discourse parsing. Many neural network models have been proposed to tackle this problem. However, the comparison for this task is not unified, so we could hardly draw clear conclusions about the effectiveness of various architectures. Here, we propose neural network models that are based on feedforward and long-short term memory architecture and systematically study the effects of varying structures. To our surprise, the best-configured feedforward architecture outperforms LSTM-based model in most cases despite thorough tuning. Further, we compare our best feedforward system with competitive convolutional and recurrent networks and find that feedforward can actually be more effective. For the first time for this task, we compile and publish outputs from previous neural and non-neural systems to establish the standard for further comparison.},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

• https://journals.uic.edu/ojs/index.php/dad/article/view/10777
• https://doi.org/10.5087/dad.2017.201

@article{Vermeul2017,
title = {On temporality in discourse annotation: Theoretical and practical considerations},
author = {Jacqueline Evers-Vermeul and Jet Hoek and Merel Scholman},
url = {https://journals.uic.edu/ojs/index.php/dad/article/view/10777},
doi = {https://doi.org/10.5087/dad.2017.201},
year = {2017},
date = {2017},
journal = {Dialogue and Discourse},
pages = {1-20},
volume = {8},
number = {2},
abstract = {

Copy BibTeX to Clipboard

We present a data-driven approach to study language use over time according to social variables (henceforth SV), considering also interactions between different variables. Besides sociolinguistic studies on language variation according to SVs (e.g., Weinreich et al. 1968, Bernstein 1971, Eckert 1989, Milroy and Milroy 1985), recently computational approaches have gained prominence (see e.g., Eisenstein 2015, Danescu-Niculescu-Mizil et al. 2013, and Nguyen et al. 2017 for an overview), not least due to an increase in data availability based on social media and an increasing awareness of the importance of linguistic variation according to SVs in the NLP community.

• clavier2017_slingpro_accepted     (0.02MB)
• https://stefaniadegaetano.files.wordpress.com/2017/07/clavier2017_slingpro_accepted.pdf

@inproceedings{Degaetano-Ortlieb2017b,
title = {Variation in language use across social variables: a data-driven approach},
author = {Stefania Degaetano-Ortlieb},
url = {https://stefaniadegaetano.files.wordpress.com/2017/07/clavier2017_slingpro_accepted.pdf},
year = {2017},
date = {2017},
booktitle = {Proceedings of the Corpus and Language Variation in English Research Conference (CLAVIER)},
address = {Bari, Italy},
abstract = {We present a data-driven approach to study language use over time according to social variables (henceforth SV), considering also interactions between different variables. Besides sociolinguistic studies on language variation according to SVs (e.g., Weinreich et al. 1968, Bernstein 1971, Eckert 1989, Milroy and Milroy 1985), recently computational approaches have gained prominence (see e.g., Eisenstein 2015, Danescu-Niculescu-Mizil et al. 2013, and Nguyen et al. 2017 for an overview), not least due to an increase in data availability based on social media and an increasing awareness of the importance of linguistic variation according to SVs in the NLP community.},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

We present an empirical approach to analyze the course of usage change in scientific writing. A great amount of linguistic research has dealt with grammatical changes, showing their gradual course of change, which nearly always progresses stepwise (see e.g. Bybee et al. 1994, Hopper and Traugott 2003, Lee 2011, De Smet and Van de Velde 2013). Less well understood is under which conditions these changes occur. According to De Smet (2016), specific expressions increase in frequency in one grammatical context, adopting a more conventionalized use, which in turn makes them available in closely related grammatical contexts.

• clavier2017-degaetano-etal_accepted_final     (0.05MB)
• https://stefaniadegaetano.files.wordpress.com/2017/07/clavier2017-degaetano-etal_accepted_final.pdf

@inproceedings{Degaetano-Ortlieb2017b,
title = {The course of grammatical change in scientific writing: Interdependency between convention and productivity},
author = {Stefania Degaetano-Ortlieb and Katrin Menzel and Elke Teich},
url = {https://stefaniadegaetano.files.wordpress.com/2017/07/clavier2017-degaetano-etal_accepted_final.pdf},
year = {2017},
date = {2017},
booktitle = {Proceedings of the Corpus and Language Variation in English Research Conference (CLAVIER)},
address = {Bari, Italy},
abstract = {We present an empirical approach to analyze the course of usage change in scientific writing. A great amount of linguistic research has dealt with grammatical changes, showing their gradual course of change, which nearly always progresses stepwise (see e.g. Bybee et al. 1994, Hopper and Traugott 2003, Lee 2011, De Smet and Van de Velde 2013). Less well understood is under which conditions these changes occur. According to De Smet (2016), specific expressions increase in frequency in one grammatical context, adopting a more conventionalized use, which in turn makes them available in closely related grammatical contexts.},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

• https://www.researchgate.net/publication/321776056_The_diachronic_development_of_combining_forms_in_scientific_writing

@article{Menzel2017,
title = {The diachronic development of combining forms in scientific writing},
author = {Katrin Menzel and Stefania Degaetano-Ortlieb},
url = {https://www.researchgate.net/publication/321776056_The_diachronic_development_of_combining_forms_in_scientific_writing},
year = {2017},
date = {2017},
journal = {Lege Artis. Language yesterday, today, tomorrow. The Journal of University of SS Cyril and Methodius in Trnava. Warsaw: De Gruyter Open},
pages = {185-249},
volume = {2},
number = {2},
abstract = {

Copy BibTeX to Clipboard

@inproceedings{Degaetano-Ortlieb2017b,
title = {An information-theoretic account on the diachronic development of discourse connectors in scientific writing},
author = {Stefania Degaetano-Ortlieb and Stefan Fischer and Vera Demberg and Elke Teich},
year = {2017},
date = {2017},
publisher = {39th DGfS AG1},
address = {Saarbr{\"u}cken, Germany},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

• W17-0503     (0.3MB)
• https://www.researchgate.net/publication/331648134_The_Making_of_the_Royal_Society_Corpus

@inproceedings{Knappen2017,
title = {The making of the Royal Society Corpus},
author = {J{\"o}rg Knappen and Stefan Fischer and Hannah Kermes and Elke Teich and Peter Fankhauser},
url = {https://www.researchgate.net/publication/331648134_The_Making_of_the_Royal_Society_Corpus},
year = {2017},
date = {2017},
booktitle = {21st Nordic Conference on Computational Linguistics (NoDaLiDa) Workshop on Processing Historical language},
pages = {7-11},
publisher = {Workshop on Processing Historical language},
address = {Gothenburg, Sweden},
abstract = {

Copy BibTeX to Clipboard

We present an approach to investigate the differences between lexical words and function words and the respective parts-of-speech from an information-theoretical point of view (cf. Shannon, 1949). We use average surprisal (AvS) to measure the amount of information transmitted by a linguistic unit. We expect to find function words to be more predictable (having a lower AvS) and lexical words to be less predictable (having a higher AvS). We also assume that function words‘ AvS is fairly constant over time and registers, while AvS of lexical words is more variable depending on time and register.

• https://www.birmingham.ac.uk/Documents/college-artslaw/corpus/conference-archives/2017/general/paper207.pdf

@inproceedings{Kermes2017,
title = {Average surprisal of parts-of-speech},
author = {Hannah Kermes and Elke Teich},
url = {https://www.birmingham.ac.uk/Documents/college-artslaw/corpus/conference-archives/2017/general/paper207.pdf},
year = {2017},
date = {2017},
publisher = {Corpus Linguistics 2017},
address = {Birmingham, UK},
abstract = {We present an approach to investigate the differences between lexical words and function words and the respective parts-of-speech from an information-theoretical point of view (cf. Shannon, 1949). We use average surprisal (AvS) to measure the amount of information transmitted by a linguistic unit. We expect to find function words to be more predictable (having a lower AvS) and lexical words to be less predictable (having a higher AvS). We also assume that function words' AvS is fairly constant over time and registers, while AvS of lexical words is more variable depending on time and register.},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

We present a data-driven approach to investigate intra-textual variation by combining entropy and surprisal. With this approach we detect linguistic variation based on phrasal lexico-grammatical patterns across sections of research articles. Entropy is used to detect patterns typical of specific sections. Surprisal is used to differentiate between more and less informationally-loaded patterns as well as type of information (topical vs. stylistic). While we here focus on research articles in biology/genetics, the methodology is especially interesting for digital humanities scholars, as it can be applied to any text type or domain and combined with additional variables (e.g. time, author or social group).

• https://aclanthology.org/W17-2209
• https://doi.org/10.18653/v1/W17-2209

@inproceedings{Degaetano-Ortlieb2017,
title = {Modeling intra-textual variation with entropy and surprisal: Topical vs. stylistic patterns},
author = {Stefania Degaetano-Ortlieb and Elke Teich},
url = {https://aclanthology.org/W17-2209},
doi = {https://doi.org/10.18653/v1/W17-2209},
year = {2017},
date = {2017},
booktitle = {Proceedings of the Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature},
pages = {68-77},
publisher = {Association for Computational Linguistics},
address = {Vancouver, Canada},
abstract = {We present a data-driven approach to investigate intra-textual variation by combining entropy and surprisal. With this approach we detect linguistic variation based on phrasal lexico-grammatical patterns across sections of research articles. Entropy is used to detect patterns typical of specific sections. Surprisal is used to differentiate between more and less informationally-loaded patterns as well as type of information (topical vs. stylistic). While we here focus on research articles in biology/genetics, the methodology is especially interesting for digital humanities scholars, as it can be applied to any text type or domain and combined with additional variables (e.g. time, author or social group).},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

1. Does following a gaze cue influence the cognitive load required for processing the corresponding linguistic referent?
2. Is considering the gaze cue costly? Is there a distribution of cognitive load between the cue and the referent?
3. Can a gaze cue have a disruptive effect on processing the linguistic referent?

• Sekicki_Staudte_CUNY2017 (1)     (2.54MB)

@miscellaneous{Sekicki2017,
title = {Cognitive load in the visual world: The facilitatory effect of gaze},
author = {Mirjana Sekicki and Maria Staudte},
year = {2017},
date = {2017},
publisher = {39th Annual Meeting of the Cognitive Science Society},
address = {London, UK},
abstract = {

1. Does following a gaze cue influence the cognitive load required for processing the corresponding linguistic referent?
2. Is considering the gaze cue costly? Is there a distribution of cognitive load between the cue and the referent?
3. Can a gaze cue have a disruptive effect on processing the linguistic referent?

Copy BibTeX to Clipboard

According to the Uniform Information Density hypothesis (UID; Jaeger 2010, inter alia), speakers strive to distribute information equally over their utterances. They do this to avoid both peaks and troughs in information density, which may lead to processing difficulty for the listener. Several studies have shown how speakers consistently make linguistic choices that result in a more equal distribution of information (e.g., Jaeger 2010, Mahowald, Fedorenko, Piantadosi, & Gibson 2013, Piantadosi, Tily, & Gibson 2011). However, it is not clear whether speakers also adapt the information density of their utterances to the processing capacity of a specific addressee. For example, when the addressee is involved in a difficult task that is clearly reducing his cognitive capacity for processing linguistic information, will the speaker lower the overall information density of her utterances to accommodate the reduced processing capacity?

• vogels     (0.06MB)

@inproceedings{Vogels2017,
title = {Referential overspecification in response to the listener's cognitive load},
author = {Jorrig Vogels and David M. Howcroft and Vera Demberg},
year = {2017},
date = {2017},
publisher = {International Cognitive Linguistics Conference},
address = {Tarttu, Estonia},
abstract = {According to the Uniform Information Density hypothesis (UID; Jaeger 2010, inter alia), speakers strive to distribute information equally over their utterances. They do this to avoid both peaks and troughs in information density, which may lead to processing difficulty for the listener. Several studies have shown how speakers consistently make linguistic choices that result in a more equal distribution of information (e.g., Jaeger 2010, Mahowald, Fedorenko, Piantadosi, & Gibson 2013, Piantadosi, Tily, & Gibson 2011). However, it is not clear whether speakers also adapt the information density of their utterances to the processing capacity of a specific addressee. For example, when the addressee is involved in a difficult task that is clearly reducing his cognitive capacity for processing linguistic information, will the speaker lower the overall information density of her utterances to accommodate the reduced processing capacity?},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

@inproceedings{Häuser2017,
title = {Age-differences in recovery from prediction error: Evidence from a simulated driving and combined sentence verification task.},
author = {Katja H{\"a}user and Vera Demberg and Jutta Kray},
year = {2017},
date = {2017-10-17},
publisher = {39th Annual Meeting of the Cognitive Science Society},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

Natural language generation (NLG) systems rely on corpora for both hand-crafted approaches in a traditional NLG architecture and for statistical end-to-end (learned) generation systems. Limitations in existing resources, however, make it difficult to develop systems which can vary the linguistic properties of an utterance as needed. For example, when users’ attention is split between a linguistic and a secondary task such as driving, a generation system may need to reduce the information density of an utterance to compensate for the reduction in user attention. We introduce a new corpus in the restaurant recommendation and comparison domain, collected in a paraphrasing paradigm, where subjects wrote texts targeting either a general audience or an elderly family member. This design resulted in a corpus of more than 5000 texts which exhibit a variety of lexical and syntactic choices and differ with respect to average word & sentence length and surprisal. The corpus includes two levels of meaning representation: flat ‘semantic stacks’ for propositional content and Rhetorical Structure Theory (RST) relations between these propositions.

• 2017-08_interspeech_extended-sparky-restaurant-corpus     (0.13MB)
• http://dx.doi.org/10.21437/Interspeech.2017-1555
• https://doi.org/10.21437/Interspeech.2017-1555

@inproceedings{Howcroft2017b,
title = {The Extended SPaRKy Restaurant Corpus: Designing a Corpus with Variable Information Density},
author = {David M. Howcroft and Dietrich Klakow and Vera Demberg},
url = {http://dx.doi.org/10.21437/Interspeech.2017-1555},
doi = {https://doi.org/10.21437/Interspeech.2017-1555},
year = {2017},
date = {2017-10-17},
booktitle = {Proc. Interspeech 2017},
pages = {3757-3761},
abstract = {Natural language generation (NLG) systems rely on corpora for both hand-crafted approaches in a traditional NLG architecture and for statistical end-to-end (learned) generation systems. Limitations in existing resources, however, make it difficult to develop systems which can vary the linguistic properties of an utterance as needed. For example, when users’ attention is split between a linguistic and a secondary task such as driving, a generation system may need to reduce the information density of an utterance to compensate for the reduction in user attention. We introduce a new corpus in the restaurant recommendation and comparison domain, collected in a paraphrasing paradigm, where subjects wrote texts targeting either a general audience or an elderly family member. This design resulted in a corpus of more than 5000 texts which exhibit a variety of lexical and syntactic choices and differ with respect to average word & sentence length and surprisal. The corpus includes two levels of meaning representation: flat ‘semantic stacks’ for propositional content and Rhetorical Structure Theory (RST) relations between these propositions.},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

Corpora of referring expressions elicited from human participants in a controlled environment are an important resource for research on automatic referring expression generation. We here present G-TUNA, a new corpus of referring expressions for German. Using the furniture stimuli set developed for the TUNA and D-TUNA corpora, our corpus extends on these corpora by providing data collected in a simulated driving dual-task setting, and additionally provides exact duration annotations for the spoken referring expressions. This corpus will hence allow researchers to analyze the interaction between referring expression length and speech rate, under conditions where the listener is under high vs. low cognitive load.

• http://www.aclweb.org/anthology/W17-3522
• https://doi.org/10.18653/v1/W17-3522

@inproceedings{W17-3522,
title = {G-TUNA: a corpus of referring expressions in German, including duration information},
author = {David M. Howcroft and Jorrig Vogels and Vera Demberg},
url = {http://www.aclweb.org/anthology/W17-3522},
doi = {https://doi.org/10.18653/v1/W17-3522},
year = {2017},
date = {2017},
booktitle = {Proceedings of the 10th International Conference on Natural Language Generation},
pages = {149-153},
publisher = {Association for Computational Linguistics},
address = {Santiago de Compostela, Spain},
abstract = {Corpora of referring expressions elicited from human participants in a controlled environment are an important resource for research on automatic referring expression generation. We here present G-TUNA, a new corpus of referring expressions for German. Using the furniture stimuli set developed for the TUNA and D-TUNA corpora, our corpus extends on these corpora by providing data collected in a simulated driving dual-task setting, and additionally provides exact duration annotations for the spoken referring expressions. This corpus will hence allow researchers to analyze the interaction between referring expression length and speech rate, under conditions where the listener is under high vs. low cognitive load.},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard

While previous research on readability has typically focused on document-level measures, recent work in areas such as natural language generation has pointed out the need of sentence-level readability measures. Much of psycholinguistics has focused for many years on processing measures that provide difficulty estimates on a word-by-word basis. However, these psycholinguistic measures have not yet been tested on sentence readability ranking tasks. In this paper, we use four psycholinguistic measures: idea density, surprisal, integration cost, and embedding depth to test whether these features are predictive of readability levels. We find that psycholinguistic features significantly improve performance by up to 3 percentage points over a standard document-level readability metric baseline.

• http://www.aclweb.org/anthology/E17-1090

@inproceedings{Howcroft2017,
title = {Psycholinguistic Models of Sentence Processing Improve Sentence Readability Ranking},
author = {David M. Howcroft and Vera Demberg},
url = {http://www.aclweb.org/anthology/E17-1090},
year = {2017},
date = {2017-10-17},
booktitle = {Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers},
pages = {958-968},
publisher = {Association for Computational Linguistics},
address = {Valencia, Spain},
abstract = {While previous research on readability has typically focused on document-level measures, recent work in areas such as natural language generation has pointed out the need of sentence-level readability measures. Much of psycholinguistics has focused for many years on processing measures that provide difficulty estimates on a word-by-word basis. However, these psycholinguistic measures have not yet been tested on sentence readability ranking tasks. In this paper, we use four psycholinguistic measures: idea density, surprisal, integration cost, and embedding depth to test whether these features are predictive of readability levels. We find that psycholinguistic features significantly improve performance by up to 3 percentage points over a standard document-level readability metric baseline.},
pubstate = {published},
type = {inproceedings}
}

Copy BibTeX to Clipboard