Papers by Goran Glavaš
Can Demographic Factors Improve Text Classification? Revisiting Demographic Adaptation in the Age of Transformers (2023.findings-eacl)
Copied to clipboard
| Challenge: | Existing studies show that incorporating demographic factors in language representations improves performance on downstream NLP tasks. |
| Approach: | They use continuous language modeling and dynamic multi-task learning to adapt pre-trained Transformers to incorporate demographic information into their representations. |
| Outcome: | The proposed model shows that the results are consistent with previous studies. |
Probing Cross-Lingual Lexical Knowledge from Multilingual Sentence Encoders (2023.eacl-main)
Copied to clipboard
| Challenge: | Pretrained multilingual language models (LMs) can be 'rewired' into effective multilingual sentence encoders (SEs) however, it remains unclear how to best leverage them to represent sub-sentence lexical items in cross-lingual lexicals. |
| Approach: | They propose a method for exposing cross-lingual lexical knowledge by additional fine-tuning through inexpensive contrastive learning that requires only a small amount of word translation pairs. |
| Outcome: | The proposed method exposes cross-lingual lexical knowledge by additional fine-tuning through inexpensive contrastive learning that requires only a small amount of word translation pairs. |
Do We Really Need Fully Unsupervised Cross-Lingual Embeddings? (D19-1)
Copied to clipboard
| Challenge: | a series of bilingual lexicon induction experiments with 15 diverse languages (210 language pairs) show that fully unsupervised CLWE methods fail for a large number of language pairs. |
| Approach: | They propose to use fully unsupervised approaches to project monolingual embeddings into a shared cross-lingual space without any cross-linguistic signal. |
| Outcome: | The proposed methods fail for a large number of language pairs, but never surpass weakly supervised methods. |
ReCoVeR the Target Language: Language Steering without Sacrificing Task Performance (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models exhibit more language confusion as they become multilingual . authors propose a lightweight approach for reducing language confusion based on language-specific steering vectors . |
| Approach: | They propose a lightweight approach to reduce language confusion by using language-specific steering vectors. |
| Outcome: | The proposed approach reduces language confusion in large language models . it leverages language-specific steering vectors for effective LLM steering . |
Babel-ImageNet: Massively Multilingual Evaluation of Vision-and-Language Representations (2024.acl-long)
Copied to clipboard
| Challenge: | Vision-and-language models with separate encoders for each modality are limited in availability. |
| Approach: | They propose a multilingual benchmark that offers (partial) translations of ImageNet labels to 100 languages, built without machine translation or manual annotation. |
| Outcome: | The proposed model outperforms models on English and low-resource languages. |
XCOPA: A Multilingual Dataset for Causal Commonsense Reasoning (2020.emnlp-main)
Copied to clipboard
| Challenge: | XCOPA dataset provides a typologically diverse dataset for commonsense reasoning in 11 languages . current methods for evaluating commonsensible reasoning in resource-poor languages are weak compared to translation-based transfer. |
| Approach: | They propose a typologically diverse multilingual dataset for causal commonsense reasoning in 11 languages. |
| Outcome: | The proposed model performs better than current methods on a resource-poor dataset compared to translation-based transfer in the 11 languages studied . |
SEAGLE: A Platform for Comparative Evaluation of Semantic Encoders for Information Retrieval (D19-3)
Copied to clipboard
| Challenge: | Existing semantic text encoding models are limited in coverage and few attempts to empirically compare them on IR tasks have been made. |
| Approach: | They propose to implement word embedding aggregators and pretrained semantic encoders and to allow for their comparative evaluation on arbitrary IR collections. |
| Outcome: | The proposed model can be exploited via an easy-to-use web interface and its modular backend (micro-service architecture) can easily be extended with additional semantic search models. |
Classification-Based Self-Learning for Weakly Supervised Bilingual Lexicon Induction (2020.acl-main)
Copied to clipboard
| Challenge: | Effective projection-based cross-lingual word embedding induction relies on the iterative self-learning procedure. |
| Approach: | They propose a classification-based approach to self-learning that allows for integration of diverse features into the iterative process. |
| Outcome: | The proposed method improves bilingual lexicon induction on a weakly supervised setup with 28 language pairs. |
Mind Your Special Tokens! On the Importance of Dedicated Sequence-End Tokens in Vision-Language Embedding Models (2026.eacl-short)
Copied to clipboard
| Challenge: | Large Vision-Language Models (LVLMs) are highly sensitive to end-of-input artifacts in fine-tuning and inference data, e.g., whether input sequences end with punctuation or newline characters. |
| Approach: | They propose to convert generative LVLMs into vision-language encoders via contrastive learning objectives and use supervised contrastive objectives to train them. |
| Outcome: | The proposed approach improves visual and text representations and improves retrieval and (semantic) similarity tasks. |
Sustainable Modular Debiasing of Language Models (2021.findings-emnlp)
Copied to clipboard
| Challenge: | Existing debiasing methods modify all of the PLM parameters, which is costly and leads to (catastrophic) forgetting of useful language knowledge. |
| Approach: | They propose a modular debiasing approach based on dedicated adapters that inject adapter modules into the original PLM layers and update only the adapters. |
| Outcome: | The proposed approach is based on dedicated adapters and retains fairness even after large-scale training. |
BenchIE: A Framework for Multi-Faceted Fact-Based Open Information Extraction Evaluation (2022.acl-long)
Copied to clipboard
| Challenge: | Existing benchmarks for OIE are incomplete and do not include all acceptable variants of the same fact. |
| Approach: | They introduce BenchIE: a benchmark and evaluation framework for comprehensive evaluation of OIE systems for English, Chinese, and German. |
| Outcome: | The proposed framework is based on fact synsets, clusters, and standardized benchmarks. |
Climbing the Tower of Treebanks: Improving Low-Resource Dependency Parsing via Hierarchical Source Selection (2021.findings-acl)
Copied to clipboard
| Challenge: | Recent work on multilingual dependency parsing focused on developing highly multilingual parsers . a recent major paradigm shift in NLP towards largescale pretrained language models has reduced the downstream relevance of supervised syntactic parse. |
| Approach: | They propose a heuristic approach to multilingual dependency parsing that heurs out the "one model to rule them all" approach by hierarchically clustering all Universal Dependencies languages based on their syntactic similarity . |
| Outcome: | The proposed approach outperforms a "one model to rule them all" approach with a heuristic selection of languages and treebanks for a target language. |
Scientia Potentia Est—On the Role of Knowledge in Computational Argumentation (2022.tacl-1)
Copied to clipboard
| Challenge: | Existing research on argumentation models does not provide a systematic overview of the types of knowledge required in CA tasks. |
| Approach: | They propose a taxonomy of the types of knowledge required in CA tasks . authors propose exploitation of these knowledge types for four main research areas . |
| Outcome: | The proposed taxonomy proposes a systematic overview of the types of knowledge required in CA tasks. |
Linking Surface Facts to Large-Scale Knowledge Graphs (2023.emnlp-main)
Copied to clipboard
| Challenge: | Open Information Extraction (OIE) methods extract facts in the form of triples . ambiguity of these triples hinders their downstream usage . |
| Approach: | They propose a benchmark that measures fact linking performance on a granular triple slot level . they propose to use a system that can detect out-of-KG entities and predicates . |
| Outcome: | The proposed benchmark can measure fact linking performance on a granular triple slot level while also measuring if a system can recognize that a surface form has no match in the existing KG. |
How Much Do LLMs Hallucinate across Languages? On Realistic Multilingual Estimation of LLM Hallucination (2025.emnlp-main)
Copied to clipboard
| Challenge: | despite LLMs becoming increasingly multilingual, most studies on detecting and quantifying LLM hallucination are English-centric . |
| Approach: | They train a multilingual hallucination detection model and conduct a large-scale study across 30 languages and 6 open-source LLM families. |
| Outcome: | The proposed model is based on an English-centric model and annotates gold data for five high-resource languages. |
Parameter-Efficient Neural Reranking for Cross-Lingual and Multilingual Retrieval (2022.coling-1)
Copied to clipboard
| Challenge: | State-of-the-art neural rankers are notoriously data-hungry and rarely used in multilingual and cross-lingual retrieval settings. |
| Approach: | They propose to use Sparse Fine-Tuning Masks and Adapters to transfer rankers trained on English data to other languages and cross-lingual setups by means of multilingual encoders. |
| Outcome: | The proposed methods outperform standard zero-shot transfer with full MMT fine-tuning while being more modular and reducing training times. |
Kardeş-NLU: Transfer to Low-Resource Languages with Big Brother’s Help – A Benchmark and Evaluation for Turkic Languages (2024.eacl-long)
Copied to clipboard
| Challenge: | Cross-lingual transfer (XLT) driven by massively multilingual language models (mmLMs) has been shown to be ineffective for low-resource (LR) target languages with little (or no) representation in mmLM’s pretraining . |
| Approach: | They propose a benchmark to evaluate cross-lingual transfer (XLT) to LR languages that do have a close HR relative and a framework to integrate Turkish into XLT. |
| Outcome: | The proposed configuration is of practical relevance for more of the world’s languages: XLT to LR languages that do have a close HR relative. |
BABELEDITS: A Benchmark and a Modular Approach for Robust Cross-lingual Knowledge Editing of Large Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for cross-lingual knowledge editing are limited in their effectiveness and robustness. |
| Approach: | They propose a new CKE benchmark that accounts for the rich variety of entity aliases within and across languages. |
| Outcome: | The proposed method is more effective than state-of-the-art methods and robust against model collapse when subjected to multiple edits. |
Is Supervised Syntactic Parsing Beneficial for Language Understanding Tasks? An Empirical Investigation (2021.eacl-main)
Copied to clipboard
| Challenge: | Traditional NLP has long held (supervised) syntactic parsing necessary for successful higher-level semantic language understanding (LU). |
| Approach: | They empirically examine the usefulness of supervised parsing for semantic LU in LM-pretrained transformer networks. |
| Outcome: | The proposed model is based on LM-pretrained transformer networks with a biaffine parsing head and fine-tuned for LU tasks. |
TransAlign: Machine Translation Encoders are Strong Word Aligners, Too (2025.findings-emnlp)
Copied to clipboard
| Challenge: | translation-based approaches to cross-lingual transfer (XLT) are limited. |
| Approach: | They propose a word aligner that utilizes the encoder of a massively multilingual MT model. |
| Outcome: | The proposed word aligner outperforms existing WA and state-of-the-art non-WA-based methods in token classification tasks. |
Multi2WOZ: A Robust Multilingual Dataset and Conversational Pretraining for Task-Oriented Dialog (2022.naacl-main)
Copied to clipboard
| Challenge: | Task-oriented dialog (TOD) is arguably one of the most popular natural language processing (NLP) application areas. |
| Approach: | They propose a multilingual multi-domain TOD dataset that spans four languages . they use a framework for multilingual conversational specialization of pretrained language models . |
| Outcome: | The proposed datasets show that they perform better than existing datasets in English . the proposed framework allows for sample-efficient few-shot transfer for TOD tasks . |
The Devil is in the Details: On Models and Training Regimes for Few-Shot Intent Classification (2023.eacl-main)
Copied to clipboard
| Challenge: | Recent methods for task-oriented dialog (ToD) intent classification use pretrained language models . but lack of informative ablations prevents identification of factors that drive performance . |
| Approach: | They propose a framework to evaluate components of Few-Shot Intent Classification . they propose to combine cross-encoder architecture and episodic meta-learning . |
| Outcome: | The proposed framework evaluates cross-encoder architecture and episodic meta-learning . it also shows that splitting episodes into support and query sets outperforms non-episodic counterparts. |
The Devil Is in the Word Alignment Details: On Translation-Based Cross-Lingual Transfer for Token Classification Tasks (2025.findings-acl)
Copied to clipboard
| Challenge: | Translation-based strategies for cross-lingual transfer XLT include label projection . word aligners (WAs) are commonly used for label projection, but low-level design decisions for using them have not been investigated . |
| Approach: | They revisit word aligners (WAs) for label projection and propose a new projection strategy that outperforms WAs. |
| Outcome: | The proposed projection strategy outperforms marker-based methods in token classification tasks. |
Don’t Stop Fine-Tuning: On Training Regimes for Few-Shot Cross-Lingual Transfer with Multilingual Language Models (2022.emnlp-main)
Copied to clipboard
| Challenge: | Recent work highlights the fallacies of zero-shot cross-lingual transfer with large multilingual models. |
| Approach: | They propose to replace sequential fine-tuning with joint fine-uning on source and target language instances. |
| Outcome: | The proposed techniques yield improved and more stable FS-XLT across the board. |
Verb Knowledge Injection for Multilingual Event Processing (2021.acl-long)
Copied to clipboard
| Challenge: | Recent studies have shown that pretrainers implicitly extract a non-negligible amount of linguistic knowledge from text corpora in an unsupervised fashion. |
| Approach: | They propose to inject explicit verb knowledge into dedicated adapter modules to complement the linguistic knowledge obtained during LM-pretraining. |
| Outcome: | The proposed model improves in English event extraction tasks, while injecting verb knowledge improves other languages. |
On Synthesizing Data for Context Attribution in Question Answering (2025.acl-long)
Copied to clipboard
Gorjan Radevski, Kiril Gashteovski, Shahbaz Syed, Christopher Malon, Sebastien Nicolas, Chia-Chien Hung, Timo Sztyler, Verena Heußer, Wiem Ben Rim, Masafumi Enomoto, Kunihiro Takeoka, Masafumi Oyamada, Goran Glavaš, Carolin Lawrence
| Challenge: | Large Language Models (LLMs) have a tendency to hallucinate, resulting in false or misleading answers. |
| Approach: | They propose a novel generative strategy for synthesizing context attribution data. |
| Outcome: | The proposed approach is highly effective for fine-tuning small LMs for context attribution in different QA tasks and domains. |
XHate-999: Analyzing and Detecting Abusive Language Across Domains and Languages (2020.coling-main)
Copied to clipboard
| Challenge: | XHate-999 is a multi-domain and multilingual evaluation data set for abusive language detection . we show that domain- and language-adaption can lead to substantially improved abusive language detecting in the target language . |
| Approach: | They propose a multi-domain and multilingual evaluation data set for abusive language detection that allows for disentanglement of domain transfer and language transfer effects. |
| Outcome: | The proposed model can significantly improve abusive language detection in the target language in the zero-shot transfer setups. |
RedditBias: A Real-World Resource for Bias Evaluation and Debiasing of Conversational Language Models (2021.acl-long)
Copied to clipboard
| Challenge: | Recent work has focused on measuring and mitigating bias in pretrained language models. |
| Approach: | They propose a dataset that measures and mitigates bias across gender,race, religion, and queerness . they compare REDDITBIAS to a widely used conversational DialoGPT model . |
| Outcome: | The proposed framework measures and mitigates bias across gender,race, religion, and queerness dimensions. |
Post-Specialisation: Retrofitting Vectors of Words Unseen in Lexical Resources (N18-1)
Copied to clipboard
| Challenge: | Word vector specialisation is a portable, light-weight approach to fine-tuning distributional word vector spaces by injecting external knowledge from rich lexical resources such as WordNet. |
| Approach: | They propose a constraint-driven vector space specialisation method that embeds external knowledge into lexical resources into a deep neural network to specialise unseen words. |
| Outcome: | The proposed method preserves useful linguistic knowledge for seen words while propagating external signal to unseen words to improve their vector representations. |
Investigating the Role of Argumentation in the Rhetorical Analysis of Scientific Publications with Neural Multi-Task Learning Models (D18-1)
Copied to clipboard
| Challenge: | Scientific publications are argumentative and often adhere to well-trodden rhetorical patterns and argumentation schemes. |
| Approach: | They investigate the link between scientific publications and rhetorical aspects such as discourse categories or citation contexts by coupling rhetorical classifiers with extraction of argumentative components. |
| Outcome: | The proposed models show significant performance gains for different rhetorical analysis tasks. |
Cross-lingual Semantic Specialization via Lexical Relation Induction (D19-1)
Copied to clipboard
| Challenge: | Semantic specialization is not available in many languages because of their incomplete or non-existent structure. |
| Approach: | They propose a method that transfers specialization from a resource-rich source language to virtually any target language. |
| Outcome: | The proposed method performs lexical simplification, dialog state tracking, and textual similarity tasks in 5 languages. |
SLICER: Sliced Fine-Tuning for Low-Resource Cross-Lingual Transfer for Named Entity Recognition (2022.emnlp-main)
Copied to clipboard
| Challenge: | Large multilingual models fail to successfully transfer to low-resource languages for zero-shot cross-lingual transfer . sliced fine-tuning for named entity recognition (SLICER) forces stronger token contextualization in the Transformer. |
| Approach: | They propose a simple yet highly effective approach for improving zero-shot cross-lingual transfer for named entity recognition to low-resource languages. |
| Outcome: | The proposed approach reduces decontextualization of token representations and classifiers . it yields consistent transfer gains for low-resource languages, the authors show . |
Self-Distillation for Model Stacking Unlocks Cross-Lingual NLU in 200+ Languages (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) excel on English NLU tasks, yet struggle to extend their NLU capabilities to underrepresented languages. |
| Approach: | They integrate machine translation models (MT) directly into LLM backbones via sample-efficient self-distillation. |
| Outcome: | The proposed model outperforms translation-test models on 127 low-resource languages. |
MVL-SIB: A Massively Multilingual Vision-Language Benchmark for Cross-Modal Topical Matching (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing multilingual vision-language (VL) benchmarks typically only cover a handful of languages, underscoring the need for evaluation data for low-resource languages. |
| Approach: | They propose a multilingual vision-language benchmark that evaluates cross-modal and text-only topical matching across 205 languages. |
| Outcome: | The proposed model performs better in cross-modal and text-only topical matching in lower-resource languages than the most multilingual benchmarks. |
Massively Multilingual Lexical Specialization of Multilingual Transformers (2023.acl-long)
Copied to clipboard
| Challenge: | Existing work focused on lexical specialization of monolingual PLMs with immense quantities of monolinguistic constraints, but recent work shows that pretrained language models can be rewired to produce high-quality word representations and perform type-level lexicals. |
| Approach: | They propose to expose massively multilingual transformers to multilingual lexical knowledge at scale using BabelNet as a source of multilingual and cross-lingual type-level lexicon knowledge. |
| Outcome: | The proposed method shows that pretrained language models can be rewired to produce high-quality word representations and perform type-level lexical tasks. |
MAD-G: Multilingual Adapter Generation for Efficient Cross-Lingual Transfer (2021.findings-emnlp)
Copied to clipboard
Alan Ansell, Edoardo Maria Ponti, Jonas Pfeiffer, Sebastian Ruder, Goran Glavaš, Ivan Vulić, Anna Korhonen
| Challenge: | Massively multilingual transformers (MMTs) have benefited from additional training of language-specific adapters, but this approach is not viable for the vast majority of languages due to limitations in their corpus size or compute budgets. |
| Approach: | They propose a multilingual ADapter generation approach which contextually generates language adapters from language representations based on typological features. |
| Outcome: | The proposed method improves cross-lingual transfer performance on part-of-speech tagging, dependency parsing, and named entity recognition tasks while remaining cost-effective. |
From Zero to Hero: On the Limitations of Zero-Shot Language Transfer with Multilingual Transformers (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies show that multilingual transformers are less effective in resource-lean scenarios and for distant languages. |
| Approach: | They propose to use massively multilingual transformers to pretrain languages . they show that MMTs are less effective in resource-lean scenarios and distant languages if they are pre-trained via language modeling . |
| Outcome: | The proposed model is less effective in resource-lean scenarios and for distant languages than cross-lingual word embeddings. |
Non-Linear Instance-Based Cross-Lingual Mapping for Non-Isomorphic Embedding Spaces (2020.acl-main)
Copied to clipboard
| Challenge: | InstaMap is a non-parametric model that learns a global linear projection by iterating between finding an optimal rotation of the source embedding space and moving each point along an instance-specific translation vector. |
| Approach: | They propose a non-parametric model that learns a global linear projection by iteratively finding an optimal rotation of the source embedding space and moving each point along an instance-specific translation vector. |
| Outcome: | The proposed model performs better than previous models on bilingual lexicon induction across a set of 28 diverse language pairs. |
Train Once, Use Flexibly: A Modular Framework for Multi-Aspect Neural News Recommendation (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Recent neural news recommenders (NNRs) extend content-based recommendation by aligning additional aspects between candidate news and user history or diversifying recommendations w.r.t. these aspects require retraining of the model with a modified objective. |
| Approach: | They introduce a modular framework for multi-aspect neural news recommendation that supports on-the-fly customization over individual aspects at inference time. |
| Outcome: | The proposed framework outperforms state-of-the-art NNRs on both content-based recommendation and single- and multi-aspect customization. |
Specializing Unsupervised Pretraining Models for Word-Level Semantic Similarity (2020.coling-main)
Copied to clipboard
| Challenge: | Unsupervised pretraining models encode only distributional knowledge encoded in text corpora, incorporated through language modeling objectives. |
| Approach: | They generalize a standard BERT model to a multi-task learning setting and integrate discrete knowledge on word-level semantic similarity into pretraining. |
| Outcome: | The proposed model outperforms the lexically blind “vanilla” model on several language understanding tasks. |
One For All & All For One: Bypassing Hyperparameter Tuning with Model Averaging for Cross-Lingual Transfer (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for zero-shot cross-lingual transfer are unreliable due to the lack of pretraining data. |
| Approach: | They propose to accumulatively average model snapshots from different runs into a single model. |
| Outcome: | The proposed protocol decouples performance maximization from hyperparameter tuning. |
Computational Analysis of Political Texts: Bridging Research Efforts Across Communities (P19-4)
Copied to clipboard
| Challenge: | Political scientists have developed and adopted natural language processing (NLP) methods to exploit text as an additional source of data in their analyses. |
| Approach: | This tutorial aims to provide a gentle introduction to methods and tasks related to computational analysis of political texts from both communities. |
| Outcome: | The main goal of this tutorial is to bring the two research communities closer to each other and contribute to faster and more significant developments in this interdisciplinary area. |
Vicinal Risk Minimization for Few-Shot Cross-lingual Transfer in Abusive Language Detection (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for few-shot cross-lingual transfer learning are limited in target languages due to the scarcity of resources. |
| Approach: | They propose a method which interpolates pairs of instances based on the angle of their representations and propose augmentation methods to enhance few-shot cross-lingual abusive language detection. |
| Outcome: | The proposed method improves few-shot cross-lingual abusive language detection in seven languages typologically distinct from English and three different domains. |
SQATIN: Supervised Instruction Tuning Meets Question Answering for Improved Dialogue NLU (2024.naacl-long)
Copied to clipboard
| Challenge: | Task-oriented dialogue (TOD) systems support users in execution of specific, well-defined tasks through natural language interaction. |
| Approach: | They propose a framework for dialog NLU based on instruction tuning and question-answering-based formulation of ID and VE tasks. |
| Outcome: | The proposed framework surpasses existing models in training and cross-domain transfer and significantly outperforms existing large language models in performance and inference efficiency. |
BAD-X: Bilingual Adapters Improve Zero-Shot Cross-Lingual Transfer (2022.naacl-main)
Copied to clipboard
| Challenge: | Massively multilingual Transformers (MMTs) have dominated research in multilingual NLP and cross-lingual transfer recently. |
| Approach: | They propose to learn bilingual language pair adapters (BAs) when the goal is to optimize performance for a particular source-target transfer direction. |
| Outcome: | The proposed framework improves performance in three standard downstream tasks and for the majority of low-resource languages. |
To Translate or Not to Translate: A Systematic Investigation of Translation-Based Cross-Lingual Transfer to Low-Resource Languages (2024.naacl-long)
Copied to clipboard
| Challenge: | XLT with multilingual language models is superfluous, says a new study . mBERT, XLM-R and mT5 are effective for cross-lingual transfer, authors say . |
| Approach: | They propose to use multilingual language models to improve cross-lingual transfer (XLT) they propose to add reliable translations to training data for XLT even for non-MT languages . |
| Outcome: | The proposed approaches outperform zero-shot XLT with mLMs, the authors show . the authors believe their findings warrant a broader inclusion of more robust translation-based baselines in XL research. |
Modular Sentence Encoders: Separating Language Specialization from Cross-Lingual Alignment (2025.acl-long)
Copied to clipboard
| Challenge: | Multilingual sentence encoders are often trained to map sentences from different languages into a shared semantic vector space . cross-lingual alignment training distorts optimal monolingual structure of semantic spaces of individual languages . a modular solution can be used for cross-linguistic tasks such as cross-language semantic similarity and zero-shot transfer . |
| Approach: | They propose a modular training system that embeds sentences from different languages into a shared semantic vector space. |
| Outcome: | The proposed solution achieves better performance across all tasks compared to monolithic models. |
Adversarial Propagation and Zero-Shot Cross-Lingual Transfer of Word Vector Specialization (D18-1)
Copied to clipboard
| Challenge: | Semantic specialization is a process of fine-tuning pre-trained distributional word vectors using external lexical knowledge to accentuate a particular semantic relation in the specialized vector space. |
| Approach: | They propose a method for specializing distributional word vectors using external lexical knowledge. |
| Outcome: | The proposed method improves on word similarity, dialog state tracking, and lexical simplification across three languages and on three tasks. |
How to (Properly) Evaluate Cross-Lingual Word Embeddings: On Strong Baselines, Comparative Analyses, and Some Misconceptions (P19-1)
Copied to clipboard
| Challenge: | Cross-lingual word embeddings (CLEs) are used for downstream NLP tasks . CLEs are based on bilingual lexicon induction (BLI) evaluations vary greatly, hindering ability to interpret performance and properties of different CLE models. |
| Approach: | They evaluate CLE models for a large number of language pairs on bilingual lexicon induction and three downstream tasks. |
| Outcome: | The proposed model performance is based on supervised and unsupervised models on bilingual lexicon induction and three downstream tasks. |
Compositional Steering of Large Language Models with Steering Tokens (2026.acl-long)
Copied to clipboard
| Challenge: | Existing work addresses compositional steering for a single behavior, but it's limited by computational costs and performance. |
| Approach: | They propose compositional steering tokens for multi-behavior steering by embedding individual behaviors into dedicated tokens via self-distillation. |
| Outcome: | The proposed steering tokens generalize well to unseen compositions compared to competing approaches. |
African or European Swallow? Benchmarking Large Vision-Language Models for Fine-Grained Object Classification (2024.emnlp-main)
Copied to clipboard
| Challenge: | Recent Large Vision Language Models demonstrate impressive abilities on image understanding and reasoning tasks. |
| Approach: | They propose a benchmark for fine-grained object classification that is difficult to evaluate . they benchmark 12 public LVLMs on and show CLIP models exhibit better performance . |
| Outcome: | The proposed model improves on 12 public LVLMs on image understanding and reasoning tasks. |
On the Limitations of Cross-lingual Encoders as Exposed by Reference-Free Machine Translation Evaluation (2020.acl-main)
Copied to clipboard
| Challenge: | a standard evaluation setup for supervised machine learning tasks does not hold for natural language generation tasks. |
| Approach: | They propose to use reference-free machine translation evaluation to compare source texts to system translations to find key limitations. |
| Outcome: | The proposed metrics perform poorly as semantic encoders for reference-free machine translation evaluation. |
DebIE: A Platform for Implicit and Explicit Debiasing of Word Embedding Spaces (2021.eacl-demos)
Copied to clipboard
| Challenge: | Recent research has shown that distributional word vector spaces often encode stereotypical human biases, such as racism and sexism. |
| Approach: | They propose a platform that measures and mitigates bias in word embeddings by executing two (mutually composable) debiasing models. |
| Outcome: | The proposed platform can measure and mitiga bias in word embeddings. |
Towards Instance-Level Parser Selection for Cross-Lingual Transfer of Dependency Parsers (2020.coling-main)
Copied to clipboard
| Challenge: | Existing methods of cross-lingual parser transfer focus on predicting the best parsers for a low-resource target language globally. |
| Approach: | They propose a cross-lingual parser transfer paradigm that uses instance-level parsers to predict the best parsing for a target language at treebank level. |
| Outcome: | The proposed model outperforms existing models on 13/20 and 14/20 test languages. |
Generalized Tuning of Distributional Word Vectors for Monolingual and Cross-Lingual Lexical Entailment (P19-1)
Copied to clipboard
| Challenge: | Lexical entailment (LE) is a core asymmetric lexical relation that supports tasks like taxonomy induction and text generation. |
| Approach: | They propose a generalized Lexical ENtailment model that captures a specialization function allowing for LE-tuning of the entire distributional vector space and not only the vectors of words seen in lexical constraints. |
| Outcome: | The proposed model improves on graded LE and shows 20% improvement over state-of-the-art LE detection. |
Probing Pretrained Language Models for Lexical Semantics (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies have focused on morphosyntactic, semantic, and world knowledge, but it remains unclear to what extent LMs derive lexical type-level knowledge from words in context. |
| Approach: | They propose to use multilingual and monolingual LMs to extract lexical type-level knowledge from words in context. |
| Outcome: | The proposed models perform well across six typologically diverse languages and five lexical tasks. |
Multilingual and Cross-Lingual Graded Lexical Entailment (P19-1)
Copied to clipboard
| Challenge: | a novel method for capturing graded (and binary) LE is developed for cross-lingual generalisation of lexical entailment . lexicale enlargement is a key principle behind hierarchical structure found in semantic networks . |
| Approach: | They propose a method for cross-lingual generalisation of GR-LE relation using hyperlex and a bilingual dictionary. |
| Outcome: | The proposed method outperforms current state-of-the-art on binary cross-lingual LE detection by a wide margin. |
Discriminating between Lexico-Semantic Relations with the Specialization Tensor Model (N18-2)
Copied to clipboard
| Challenge: | Existing methods to specialize distributional vectors to better reflect a particular relation are lacking in modern natural language processing. |
| Approach: | They propose a feed-forward neural architecture for discriminating between lexico-semantic relations . they propose to train relation classifiers using lexical relations from external resources . |
| Outcome: | The proposed model outperforms state-of-the-art models on two benchmark datasets and exhibits stable performance across languages. |
NewsRecLib: A PyTorch-Lightning Library for Neural News Recommendation (2023.emnlp-demo)
Copied to clipboard
| Challenge: | Existing open source repositories expose a multitude of programming languages, libraries, and implementation differences, hindering reproducibility and extensibility. |
| Approach: | NewsRecLib is an open-source library for training and evaluating neural news recommendation models. |
| Outcome: | The open-source library provides implementations of several prominent neural models, training methods, standard evaluation benchmarks, and evaluation metrics for news recommendation. |
On Generalization across Measurement Systems: LLMs Entail More Test-Time Compute for Underrepresented Cultures (2025.acl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) should be able to provide accurate information irrespective of the measurement system at hand . |
| Approach: | They use newly compiled datasets to test if this is true for seven open-source LLMs. |
| Outcome: | The proposed model can provide accurate information regardless of the measurement system at hand. |
One Script Instead of Hundreds? On Pretraining Romanized Encoder Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | a recent study has focused on setups that favor romanization for cross-lingual transfer . a fidelity-based approach is needed to improve performance for high-resource languages . |
| Approach: | They propose to pretrain LMs from scratch on romanized and original texts for six languages . they find that romanization improves encoding efficiency for segmental scripts at a negligible cost . |
| Outcome: | The proposed method reduces the loss of script-specific information and dilution of language-specific representations from increased subword overlap. |
Natural Language Processing for Multilingual Task-Oriented Dialogue (2022.acl-tutorials)
Copied to clipboard
| Challenge: | a tutorial will examine the challenges and gaps in multilingual ToD research . multilingual systems are difficult to build, and are limited to English and other languages . |
| Approach: | This tutorial will discuss the importance of multilingual task-oriented dialogue systems . it will provide an overview of current research gaps, challenges and initiatives related to multilingual ToD systems - with a particular focus on their connections to current research and challenges in multilingual and low-resource NLP. |
| Outcome: | This tutorial will provide an overview of current research gaps, challenges and initiatives related to multilingual ToD systems. |
Explicit Retrofitting of Distributional Word Vectors (P18-1)
Copied to clipboard
| Challenge: | Existing models for word vector specialization focus on word co-occurrences from large text corpora, resulting in a tendency to fuse semantic similarity with other types of semantic relatedness. |
| Approach: | They propose to transform external lexico-semantic relations into training examples and learn an explicit retrofitting model. |
| Outcome: | The proposed model can specialize vector spaces of new languages and translate them to other languages. |
AnnIE: An Annotation Platform for Constructing Complete Open Information Extraction Benchmark (2022.acl-demo)
Copied to clipboard
Niklas Friedrich, Kiril Gashteovski, Mingying Yu, Bhushan Kotnis, Carolin Lawrence, Mathias Niepert, Goran Glavaš
| Challenge: | Open Information Extraction (OIE) is the task of extracting facts from sentences in the form of relations and their corresponding arguments in schema-free manner. |
| Approach: | They propose an interactive annotation platform that facilitates annotating complete facts from input sentences. |
| Outcome: | The proposed platform facilitates such challenging annotation tasks and supports creation of fact-oriented OIE evaluation benchmarks. |
Free Lunch: Robust Cross-Lingual Transfer via Model Checkpoint Averaging (2023.acl-long)
Copied to clipboard
| Challenge: | Massively multilingual language models have shown strong performance in zero-shot (ZS-XLT) and few-shot cross-lingual transfer setups where models are fine-tuned on task data in a source language are transferred without any or with only a few annotated instances to the target language(s). |
| Approach: | They propose a method that averages different checkpoints during task fine-tuning to improve model robustness. |
| Outcome: | The proposed method overestimates model performance in cross-lingual transfer setups where models are evaluated at checkpoints that generalize best to validation instances in the target languages. |
Does Object Grounding Really Reduce Hallucination of Large Vision-Language Models? (2024.emnlp-main)
Copied to clipboard
| Challenge: | Large vision-language models (LVLMs) often hallucinate and produce captions that mention concepts that cannot be found in the image. |
| Approach: | They propose to add grounding objectives to captions that explicitly align image regions or objects to text spans to reduce hallucination. |
| Outcome: | The proposed evaluation protocol reduces the amount of hallucination in LVLMs by adding grounding objectives. |
Leveraging Open Information Extraction for More Robust Domain Transfer of Event Trigger Detection (2024.findings-eacl)
Copied to clipboard
| Challenge: | Event detection is a crucial information extraction task in many domains, such as Wikipedia or news. |
| Approach: | They propose to use subject-object relations obtained from a rule-based open information extraction system to enhance domain transfer of trigger detection (TD) they combine this enhanced transfer with masked language modeling on the target domain, observing further TD transfer gains. |
| Outcome: | The proposed model improves the transfer of triggers between domains and reduces performance drops when using a low-resource source domain to a high-res target domain. |
Centurio: On Drivers of Multilingual Ability of Large Vision-Language Model (2025.acl-long)
Copied to clipboard
Gregor Geigle, Florian Schneider, Carolin Holtermann, Chris Biemann, Radu Timofte, Anne Lauscher, Goran Glavaš
| Challenge: | Existing models for large vision-language tasks are trained on English data, which makes them struggle to understand non-English input and fail to generate output in the desired target language. |
| Approach: | They conduct multi-stage experiments on 13 vision-language tasks and 43 languages . they find that one can include as many as 100 training languages simultaneously with as little as 25-50% of non-English data . |
| Outcome: | The proposed model outperforms existing models in 14 tasks and 56 languages. |
LexFit: Lexical Fine-Tuning of Pretrained Language Models (2021.acl-long)
Copied to clipboard
| Challenge: | Transformer-based language models implicitly store a wealth of lexical semantic knowledge, but it is non-trivial to extract that knowledge effectively from their parameters. |
| Approach: | They propose to expose and enrich lexical knowledge from transformer-based language models to serve as effective decontextualized word encoders even when fed input words "in isolation" |
| Outcome: | The proposed model outperforms standard static WEs and vanilla LMs in lexical tasks over four established tasks in 8 languages. |
AdaSent: Efficient Domain-Adapted Sentence Embeddings for Few-Shot Classification (2023.emnlp-main)
Copied to clipboard
| Challenge: | Recent work has found that few-shot sentence classification based on pre-trained Sentence Encoders (SEs) is efficient, robust, and effective. |
| Approach: | They propose a domain-specialization approach that decouples SEPT from DAPT by training a SEPT adapter on a pre-trained PLM. |
| Outcome: | The proposed approach matches or surpasses the performance of full SEPT on DAPT-ed PLMs while significantly reducing training costs. |
Large Language Models are Miscalibrated In-Context Learners (2025.findings-acl)
Copied to clipboard
| Challenge: | In-context Learning and Supervised Fine-Tuning have emerged as pre-dominant methodologies for machine learning and NLP. |
| Approach: | They propose to use self-ensembling to improve both performance and calibration of language models. |
| Outcome: | The proposed learning paradigms can achieve better calibration and better performance than the previous learning paradigm. |
DS-TOD: Efficient Domain Specialization for Task-Oriented Dialog (2022.findings-acl)
Copied to clipboard
| Challenge: | Recent work shows that self-supervised dialog-specific pretraining on large conversational datasets yields substantial gains over traditional language modeling (LM) pretraining. |
| Approach: | They propose a resource-efficient and modular domain specialization by means of domain adapters in which domain knowledge is encoded. |
| Outcome: | The proposed framework extracts domain-specific terms and then uses them to build DomainCC and DomainReddit resources based on masked language modeling and response selection objectives. |