Rik van Noord

Esther Ploeger, Huiyuan Lai, Rik van Noord et al.

Machine translations are found to be lexically poorer than human translations. The loss of lexical diversity through MT poses an issue in the automatic translation of literature, where it matters not only what is written, but also how it is written. Current methods for increasing lexical diversity in MT are rigid. Yet, as we demonstrate, the degree of lexical diversity can vary considerably across different novels. Thus, rather than aiming for the rigid increase of lexical diversity, we reframe the task as recovering what is lost in the machine translation process. We propose a novel approach that consists of reranking translation candidates with a classifier that distinguishes between original and translated text. We evaluate our approach on 31 English-to-Dutch book translations, and find that, for certain books, our approach retrieves lexical diversity scores that are close to human translation.

Yun Hao, Reihaneh Amooie, Wietse de Vries et al.

Automatic speech recognition (ASR) has improved substantially in recent years, yet performance remains limited for low-resource languages. Large language models (LLMs) have shown promise for improving ASR through generative error correction (GER), but their effectiveness in low-resource settings remains underexplored. In addition, it remains unclear to what extent data contamination influences the reported improvements in LLM-based GER. This study investigates LLM-based GER for low-resource Frisian. In addition to a public corpus, we construct and use a Frisian offline dataset with non-public texts for evaluation to control for potential data contamination. Results show that GER improves ASR performance in most settings, with the best GPT-5.1 results surpassing oracle WERs. Comparable gains on the offline dataset indicate that improvements reflect true correction ability. We further provide a detailed error analysis revealing model correction patterns.

Kyo Gerrits, Rik van Noord, Ana Guerberof Arenas

This article investigates the performance of automatic evaluation metrics (AEMs) and LLM-as-a-judge evaluation on literary translation across multiple languages, genres, and translation modalities. The aim is to assess how well these tools align with professionals when evaluating translation, creativity (creative shifts & errors), and see if they can substitute laborious manual annotations. A dataset of literary translations across three modalities (human translation, machine translation, and post-editing), three genres and three language pairs was created and annotated in detail for creativity by experienced professional literary translators. The results show that both AEMs and LLM-as-a-judge evaluations correlate poorly with professional evaluations on creativity, with LLM-as-a-judge showing a systematic bias in favour of machine-translated texts and penalising creative and culturally appropriate solutions. Moreover, performance is consistently worse for more literary genres such as poetry. This highlights fundamental limitations of current automatic evaluation tools for literary translation and the need to create new tools that do not frequently consider out of routine translations as errors.

Chunliu Wang, Rik van Noord, Johan Bos

A bare meaning representation can be expressed in various ways using natural language, depending on how the information is structured on the surface level. We are interested in finding ways to control topic-focus articulation when generating text from meaning. We focus on distinguishing active and passive voice for sentences with transitive verbs. The idea is to add pragmatic information such as topic to the meaning representation, thereby forcing either active or passive voice when given to a natural language generation system. We use graph neural models because there is no explicit information about word order in a meaning represented by a graph. We try three different methods for topic-focus articulation (TFA) employing graph neural models for a meaning-to-text generation task. We propose a novel encoding strategy about node aggregation in graph neural models, which instead of traditional encoding by aggregating adjacent node information, learns node representations by using depth-first search. The results show our approach can get competitive performance with state-of-art graph models on general text generation, and lead to significant improvements on the task of active-passive conversion compared to traditional adjacency-based aggregation strategies. Different types of TFA can have a huge impact on the performance of the graph models.

Xiao Zhang, Chunliu Wang, Rik van Noord et al.

The Parallel Meaning Bank (PMB) serves as a corpus for semantic processing with a focus on semantic parsing and text generation. Currently, we witness an excellent performance of neural parsers and generators on the PMB. This might suggest that such semantic processing tasks have by and large been solved. We argue that this is not the case and that performance scores from the past on the PMB are inflated by non-optimal data splits and test sets that are too easy. In response, we introduce several changes. First, instead of the prior random split, we propose a more systematic splitting approach to improve the reliability of the standard test data. Second, except for the standard test set, we also propose two challenge sets: one with longer texts including discourse structure, and one that addresses compositional generalization. We evaluate five neural models for semantic parsing and meaning-to-text generation. Our results show that model performance declines (in some cases dramatically) on the challenge sets, revealing the limitations of neural models when confronting such challenges.

Nikola Ljubešić, Vít Suchomel, Peter Rupnik et al.

The world of language models is going through turbulent times, better and ever larger models are coming out at an unprecedented speed. However, we argue that, especially for the scientific community, encoder models of up to 1 billion parameters are still very much needed, their primary usage being in enriching large collections of data with metadata necessary for downstream research. We investigate the best way to ensure the existence of such encoder models on the set of very closely related languages - Croatian, Serbian, Bosnian and Montenegrin, by setting up a diverse benchmark for these languages, and comparing the trained-from-scratch models with the new models constructed via additional pretraining of existing multilingual models. We show that comparable performance to dedicated from-scratch models can be obtained by additionally pretraining available multilingual models even with a limited amount of computation. We also show that neighboring languages, in our case Slovenian, can be included in the additional pretraining with little to no loss in the performance of the final model.

Rik van Noord, Taja Kuzman, Peter Rupnik et al.

Large, curated, web-crawled corpora play a vital role in training language models (LMs). They form the lion's share of the training data in virtually all recent LMs, such as the well-known GPT, LLaMA and XLM-RoBERTa models. However, despite this importance, relatively little attention has been given to the quality of these corpora. In this paper, we compare four of the currently most relevant large, web-crawled corpora (CC100, MaCoCu, mC4 and OSCAR) across eleven lower-resourced European languages. Our approach is two-fold: first, we perform an intrinsic evaluation by performing a human evaluation of the quality of samples taken from different corpora; then, we assess the practical impact of the qualitative differences by training specific LMs on each of the corpora and evaluating their performance on downstream tasks. We find that there are clear differences in quality of the corpora, with MaCoCu and OSCAR obtaining the best results. However, during the extrinsic evaluation, we actually find that the CC100 corpus achieves the highest scores. We conclude that, in our experiments, the quality of the web-crawled corpora does not seem to play a significant role when training LMs.

Huiyuan Lai, Esther Ploeger, Rik van Noord et al.

Neural machine translation (NMT) systems amplify lexical biases present in their training data, leading to artificially impoverished language in output translations. These language-level characteristics render automatic translations different from text originally written in a language and human translations, which hinders their usefulness in for example creating evaluation datasets. Attempts to increase naturalness in NMT can fall short in terms of content preservation, where increased lexical diversity comes at the cost of translation accuracy. Inspired by the reinforcement learning from human feedback framework, we introduce a novel method that rewards both naturalness and content preservation. We experiment with multiple perspectives to produce more natural translations, aiming at reducing machine and human translationese. We evaluate our method on English-to-Dutch literary translation, and find that our best model produces translations that are lexically richer and exhibit more properties of human-written language, without loss in translation accuracy.

Malina Chichirau, Rik van Noord, Antonio Toral

We tackle the task of automatically discriminating between human and machine translations. As opposed to most previous work, we perform experiments in a multilingual setting, considering multiple languages and multilingual pretrained language models. We show that a classifier trained on parallel data with a single source language (in our case German-English) can still perform well on English translations that come from different source languages, even when the machine translations were produced by other systems than the one it was trained on. Additionally, we demonstrate that incorporating the source text in the input of a multilingual classifier improves (i) its accuracy and (ii) its robustness on cross-system evaluation, compared to a monolingual classifier. Furthermore, we find that using training data from multiple source languages (German, Russian, and Chinese) tends to improve the accuracy of both monolingual and multilingual classifiers. Finally, we show that bilingual classifiers and classifiers trained on multiple source languages benefit from being trained on longer text sequences, rather than on sentences.

Lasha Abzianidze, Rik van Noord, Chunliu Wang et al.

This paper gives a general description of the ideas behind the Parallel Meaning Bank, a framework with the aim to provide an easy way to annotate compositional semantics for texts written in languages other than English. The annotation procedure is semi-automatic, and comprises seven layers of linguistic information: segmentation, symbolisation, semantic tagging, word sense disambiguation, syntactic structure, thematic role labelling, and co-reference. New languages can be added to the meaning bank as long as the documents are based on translations from English, but also introduce new interesting challenges on the linguistics assumptions underlying the Parallel Meaning Bank.

Rik van Noord, Antonio Toral, Johan Bos

We combine character-level and contextual language model representations to improve performance on Discourse Representation Structure parsing. Character representations can easily be added in a sequence-to-sequence model in either one encoder or as a fully separate encoder, with improvements that are robust to different language models, languages and data sets. For English, these improvements are larger than adding individual sources of linguistic information or adding non-contextual embeddings. A new method of analysis based on semantic tags demonstrates that the character-level representations improve performance across a subset of selected semantic phenomena.

Lasha Abzianidze, Rik van Noord, Hessel Haagsma et al.

The paper presents the IWCS 2019 shared task on semantic parsing where the goal is to produce Discourse Representation Structures (DRSs) for English sentences. DRSs originate from Discourse Representation Theory and represent scoped meaning representations that capture the semantics of negation, modals, quantification, and presupposition triggers. Additionally, concepts and event-participants in DRSs are described with WordNet synsets and the thematic roles from VerbNet. To measure similarity between two DRSs, they are represented in a clausal form, i.e. as a set of tuples. Participant systems were expected to produce DRSs in this clausal form. Taking into account the rich lexical information, explicit scope marking, a high number of shared variables among clauses, and highly-constrained format of valid DRSs, all these makes the DRS parsing a challenging NLP task. The results of the shared task displayed improvements over the existing state-of-the-art parser.

Malvina Nissim, Rik van Noord, Rob van der Goot

Analogies such as "man is to king as woman is to X" are often used to illustrate the amazing power of word embeddings. Concurrently, they have also been used to expose how strongly human biases are encoded in vector spaces built on natural language, like "man is to computer programmer as woman is to homemaker". Recent work has shown that analogies are in fact not such a diagnostic for bias, and other methods have been proven to be more apt to the task. However, beside the intrinsic problems with the analogy task as a bias detection tool, in this paper we show that a series of issues related to how analogies have been implemented and used might have yielded a distorted picture of bias in word embeddings. Human biases are present in word embeddings and need to be addressed. Analogies, though, are probably not the right tool to do so. Also, the way they have been most often used has exacerbated some possibly non-existing biases and perhaps hid others. Because they are still widely popular, and some of them have become classics within and outside the NLP community, we deem it important to provide a series of clarifications that should put well-known, and potentially new cases into the right perspective.

Rik van Noord, Lasha Abzianidze, Antonio Toral et al.

Neural methods have had several recent successes in semantic parsing, though they have yet to face the challenge of producing meaning representations based on formal semantics. We present a sequence-to-sequence neural semantic parser that is able to produce Discourse Representation Structures (DRSs) for English sentences with high accuracy, outperforming traditional DRS parsers. To facilitate the learning of the output, we represent DRSs as a sequence of flat clauses and introduce a method to verify that produced DRSs are well-formed and interpretable. We compare models using characters and words as input and see (somewhat surprisingly) that the former performs better than the latter. We show that eliminating variable names from the output using De Bruijn-indices increases parser performance. Adding silver training data boosts performance even further.

Marloes Kuijper, Mike van Lenthe, Rik van Noord

The present study describes our submission to SemEval 2018 Task 1: Affect in Tweets. Our Spanish-only approach aimed to demonstrate that it is beneficial to automatically generate additional training data by (i) translating training data from other languages and (ii) applying a semi-supervised learning method. We find strong support for both approaches, with those models outperforming our regular models in all subtasks. However, creating a stepwise ensemble of different models as opposed to simply averaging did not result in an increase in performance. We placed second (EI-Reg), second (EI-Oc), fourth (V-Reg) and fifth (V-Oc) in the four Spanish subtasks we participated in.

Rik van Noord, Lasha Abzianidze, Hessel Haagsma et al.

Semantic parsing offers many opportunities to improve natural language understanding. We present a semantically annotated parallel corpus for English, German, Italian, and Dutch where sentences are aligned with scoped meaning representations in order to capture the semantics of negation, modals, quantification, and presupposition triggers. The semantic formalism is based on Discourse Representation Theory, but concepts are represented by WordNet synsets and thematic roles by VerbNet relations. Translating scoped meaning representations to sets of clauses enables us to compare them for the purpose of semantic parser evaluation and checking translations. This is done by computing precision and recall on matching clauses, in a similar way as is done for Abstract Meaning Representations. We show that our matching tool for evaluating scoped meaning representations is both accurate and efficient. Applying this matching tool to three baseline semantic parsers yields F-scores between 43% and 54%. A pilot study is performed to automatically find changes in meaning by comparing meaning representations of translations. This comparison turns out to be an additional way of (i) finding annotation mistakes and (ii) finding instances where our semantic analysis needs to be improved.

Rik van Noord, Johan Bos

We evaluate the character-level translation method for neural semantic parsing on a large corpus of sentences annotated with Abstract Meaning Representations (AMRs). Using a sequence-to-sequence model, and some trivial preprocessing and postprocessing of AMRs, we obtain a baseline accuracy of 53.1 (F-score on AMR-triples). We examine five different approaches to improve this baseline result: (i) reordering AMR branches to match the word order of the input sentence increases performance to 58.3; (ii) adding part-of-speech tags (automatically produced) to the input shows improvement as well (57.2); (iii) So does the introduction of super characters (conflating frequent sequences of characters to a single character), reaching 57.4; (iv) optimizing the training process by using pre-training and averaging a set of models increases performance to 58.7; (v) adding silver-standard training data obtained by an off-the-shelf parser yields the biggest improvement, resulting in an F-score of 64.0. Combining all five techniques leads to an F-score of 71.0 on holdout data, which is state-of-the-art in AMR parsing. This is remarkable because of the relative simplicity of the approach.

Rik van Noord, Johan Bos

We evaluate a semantic parser based on a character-based sequence-to-sequence model in the context of the SemEval-2017 shared task on semantic parsing for AMRs. With data augmentation, super characters, and POS-tagging we gain major improvements in performance compared to a baseline character-level model. Although we improve on previous character-based neural semantic parsing models, the overall accuracy is still lower than a state-of-the-art AMR parser. An ensemble combining our neural semantic parser with an existing, traditional parser, yields a small gain in performance.

Lasha Abzianidze, Johannes Bjerva, Kilian Evang et al.

The Parallel Meaning Bank is a corpus of translations annotated with shared, formal meaning representations comprising over 11 million words divided over four languages (English, German, Italian, and Dutch). Our approach is based on cross-lingual projection: automatically produced (and manually corrected) semantic annotations for English sentences are mapped onto their word-aligned translations, assuming that the translations are meaning-preserving. The semantic annotation consists of five main steps: (i) segmentation of the text in sentences and lexical items; (ii) syntactic parsing with Combinatory Categorial Grammar; (iii) universal semantic tagging; (iv) symbolization; and (v) compositional semantic analysis based on Discourse Representation Theory. These steps are performed using statistical models trained in a semi-supervised manner. The employed annotation models are all language-neutral. Our first results are promising.