Search Results for author: Torsten Zesch

Found 65 papers, 9 papers with code

Analyzing the Real Vulnerability of Hate Speech Detection Systems against Targeted Intentional Noise

no code implementations COLING (WNUT) 2022 Piush Aggarwal, Torsten Zesch

Hate speech detection systems have been shown to be vulnerable against obfuscation attacks, where a potential hater tries to circumvent detection by deliberately introducing noise in their posts.

Hate Speech Detection

Implicit Phenomena in Short-answer Scoring Data

no code implementations ACL (unimplicit) 2021 Marie Bexte, Andrea Horbach, Torsten Zesch

We therefore quantify to what extent implicit language phenomena occur in short answer datasets and examine the influence they have on automatic scoring performance.

Word Embeddings

‘Meet me at the ribary’ – Acceptability of spelling variants in free-text answers to listening comprehension prompts

no code implementations NAACL (BEA) 2022 Ronja Laarmann-Quante, Leska Schwarz, Andrea Horbach, Torsten Zesch

When listening comprehension is tested as a free-text production task, a challenge for scoring the answers is the resulting wide range of spelling variants.

A Crash Course on Ethics for Natural Language Processing

no code implementations NAACL (TeachingNLP) 2021 Annemarie Friedrich, Torsten Zesch

It is generally agreed upon in the natural language processing (NLP) community that ethics should be integrated into any curriculum.


Similarity-Based Content Scoring - How to Make S-BERT Keep Up With BERT

1 code implementation NAACL (BEA) 2022 Marie Bexte, Andrea Horbach, Torsten Zesch

The dominating paradigm for content scoring is to learn an instance-based model, i. e. to use lexical features derived from the learner answers themselves.

LeSpell - A Multi-Lingual Benchmark Corpus of Spelling Errors to Develop Spellchecking Methods for Learner Language

1 code implementation LREC 2022 Marie Bexte, Ronja Laarmann-Quante, Andrea Horbach, Torsten Zesch

Spellchecking text written by language learners is especially challenging because errors made by learners differ both quantitatively and qualitatively from errors made by already proficient learners.

C-Test Collector: A Proficiency Testing Application to Collect Training Data for C-Tests

no code implementations EACL (BEA) 2021 Christian Haring, Rene Lehmann, Andrea Horbach, Torsten Zesch

We present the C-Test Collector, a web-based tool that allows language learners to test their proficiency level using c-tests.

HateProof: Are Hateful Meme Detection Systems really Robust?

no code implementations11 Feb 2023 Piush Aggarwal, Pranit Chawla, Mithun Das, Punyajoy Saha, Binny Mathew, Torsten Zesch, Animesh Mukherjee

Empirically, we find a noticeable performance drop of as high as 10% in the macro-F1 score for certain attacks.

Contrastive Learning

Chinese Content Scoring: Open-Access Datasets and Features on Different Segmentation Levels

no code implementations Asian Chapter of the Association for Computational Linguistics 2020 Yuning Ding, Andrea Horbach, Torsten Zesch

As a review of prior work for Chinese content scoring shows a lack of open-access data in the field, we present two short-answer data sets for Chinese.

Decomposing and Comparing Meaning Relations: Paraphrasing, Textual Entailment, Contradiction, and Specificity

no code implementations LREC 2020 Venelin Kovatchev, Darina Gold, M. Antonia Marti, Maria Salamo, Torsten Zesch

We use the typology to annotate a corpus of 520 sentence pairs in English and we demonstrate that unlike previous typologies, SHARel can be applied to all relations of interest with a high inter-annotator agreement.

Natural Language Inference Specificity

A Legal Approach to Hate Speech -- Operationalizing the EU's Legal Framework against the Expression of Hatred as an NLP Task

no code implementations7 Apr 2020 Frederike Zufall, Marius Hamacher, Katharina Kloppenborg, Torsten Zesch

We propose a 'legal approach' to hate speech detection by operationalization of the decision as to whether a post is subject to criminal law into an NLP task.

Decision Making Hate Speech Detection

Divide and Extract -- Disentangling Clause Splitting and Proposition Extraction

no code implementations RANLP 2019 Darina Gold, Torsten Zesch

The resulting proposition evaluation dataset allows us to independently compare the performance of proposition extraction systems on simple and complex clauses.

Cross-Lingual Content Scoring

no code implementations WS 2018 Andrea Horbach, Sebastian Stennmanns, Torsten Zesch

We investigate the feasibility of cross-lingual content scoring, a scenario where training and test data in an automatic scoring task are from two different languages.

Machine Translation

Agree or Disagree: Predicting Judgments on Nuanced Assertions

1 code implementation SEMEVAL 2018 Michael Wojatzki, Torsten Zesch, Saif Mohammad, Svetlana Kiritchenko

Being able to predict whether people agree or disagree with an assertion (i. e. an explicit, self-contained statement) has several applications ranging from predicting how many people will like or dislike a social media post to classifying posts based on whether they are in accordance with a particular point of view.

The Influence of Spelling Errors on Content Scoring Performance

no code implementations WS 2017 Andrea Horbach, Yuning Ding, Torsten Zesch

Spelling errors occur frequently in educational settings, but their influence on automatic scoring is largely unknown.

BIG-bench Machine Learning

Fine-grained essay scoring of a complex writing task for native speakers

no code implementations WS 2017 Andrea Horbach, Dirk Scholten-Akoun, Yuning Ding, Torsten Zesch

Automatic essay scoring is nowadays successfully used even in high-stakes tests, but this is mainly limited to holistic scoring of learner essays.

Do LSTMs really work so well for PoS tagging? -- A replication study

no code implementations EMNLP 2017 Tobias Horsmann, Torsten Zesch

A recent study by Plank et al. (2016) found that LSTM-based PoS taggers considerably improve over the current state-of-the-art when evaluated on the corpora of the Universal Dependencies project that use a coarse-grained tagset.

Feature Engineering Part-Of-Speech Tagging +1

Assigning Fine-grained PoS Tags based on High-precision Coarse-grained Tagging

no code implementations COLING 2016 Tobias Horsmann, Torsten Zesch

We propose a new approach to PoS tagging where in a first step, we assign a coarse-grained tag corresponding to the main syntactic category.


Predicting the Difficulty of Language Proficiency Tests

no code implementations TACL 2014 Lisa Beinborn, Torsten Zesch, Iryna Gurevych

Language proficiency tests are used to evaluate and compare the progress of language learners.

Cannot find the paper you are looking for? You can Submit a new open access paper.