Morocco

Arabic dialects have long been under-represented in Natural Language Processing (NLP) research due to their non-standardization and high variability, which pose challenges for computational modeling. Recent advances in the field, such as Large Language Models (LLMs), offer promising avenues to address this gap by enabling Arabic to be modeled as a pluricentric language rather than a monolithic system. This paper presents Aladdin-FTI, our submission to the AMIYA shared task. The proposed system is designed to both generate and translate dialectal Arabic (DA). Specifically, the model supports text generation in Moroccan, Egyptian, Palestinian, Syrian, and Saudi dialects, as well as bidirectional translation between these dialects, Modern Standard Arabic (MSA), and English. The code and trained model will be released upon paper acceptance.

EACL 2026 Main Conference

Aladdin-FTI @ AMIYA Three Wishes for Arabic NLP: Fidelity, Diglossia, and Multidialectal Generation

workshop paper

#### *Message from the General Chair, Aline Villavicencio*
I’m delighted and honoured to welcome you to the 19th Conference of the European Chapter of the Association for Computational Linguistics (EACL 2026), taking place in the beautiful city of Rabat, in Morocco, in March 24-29, 2026. EACL is the flagship European conference of the Association and EACL 2026 proudly continues our field’s tradition of excellence in scholarship, innovation, and inclusivity. I am deeply grateful to the many volunteers whose dedication, generosity, and tireless efforts have made this conference possible.
For the first time EACL is being hosted in the African continent. This is an important milestone for our community, and we are grateful to our Moroccan hosts for enabling this historic moment by bringing this edition of EACL to Rabat. We are also delighted that the Second Arabic NLP School is co-located with EACL. We hope attendees enjoy this wonderful opportunity to strengthen ties with the Computational Linguistics communities across the African continent. *[Read full message](https://drive.google.com/file/d/14NlmHvwM6fPJuMmOvVh7K0vtQbEyv3SZ/view?usp=sharing)*<br><br>

<html><button style="display: inline-flex; align-items: center; justify-content: center; white-space: nowrap; border-radius: 9999px; font-weight: bold; background: #7c3aed; color: white; font-family: 'Space Grotesk', sans-serif; height: 40px; font-size: 16px; padding: 0 20px; border: none; cursor: pointer" onclick="window.open('https://underline.io/events/522/reception','_blank')">Go to Workshops and Tutorials Program</button></html>
<br><br><br>

To access this event page, you need to log in with the **email address you registered with**. <br>Access credentials will be sent to your email from Underline -  subject line "Welcome to EACL 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

Please log in

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://2026.eacl.org/registration/) first.

**Online Registration Form**: https://acl.swoogo.com/eacl2026

Registration Required

Welcome to the 19th Conference of the European Chapter of the Association for Computational Linguistics (EACL). EACL 2026 will be held in Rabat, Morocco, from March 24–29, 2026. 

Grammatical error correction (GEC) aims to improve text quality and readability. Previous work on the task focused primarily on high-resource languages, while low-resource languages lack robust tools. To address this shortcoming, we present a study on GEC for Zarma, a language spoken by over five million people in West Africa. We compare three approaches: rule-based methods, machine translation (MT) models, and large language models (LLMs). We evaluated GEC models using a dataset of more than 250,000 examples, including synthetic and human-annotated data. Our results showed that the MT-based approach using M2M100 outperforms others, with a detection rate of 95.82% and a suggestion accuracy of 78.90% in automatic evaluations (AE) and an average score of 3.0 out of 5.0 in manual evaluation (ME) from native speakers for grammar and logical corrections. The rule-based method was effective for spelling errors but failed on complex context-level errors. LLMs---Gemma 2b and MT5-small---showed moderate performance. Our work supports use of MT models to enhance GEC in low-resource settings, and we validated these results with Bambara, another West African language.

Grammatical Error Correction for Low-Resource Languages: The Case of Zarma

Fine-tuning multilingual models for low-resource dialect translation frequently encounters a “plausibility over faithfulness” dilemma, resulting in severe semantic drift on dialect-specific tokens. We term this phenomenon the “Probability Trap,” where models prioritize statistical fluency over semantic fidelity. To address this, we propose MVS-Rank (Multi-View Scoring Reranking), a generate-then-rerank framework that decouples evaluation from generation. Our method assesses translation candidates through three complementary perspectives: (1) Source-Side Faithfulness via a Reverse Translation Model to anchor semantic fidelity; (2) Local Fluency using Masked Language Models to ensure syntactic precision; and (3) Global Fluency leveraging Large Language Models to capture discourse coherence. Extensive experiments on Cantonese-Mandarin benchmarks demonstrate that MVS-Rank achieves state-of-the-art performance, significantly outperforming strong fine-tuning baselines by effectively rectifying hallucinations while maintaining high fluency.

Escaping the Probability Trap: Mitigating Semantic Drift in Cantonese-Mandarin Translation

We use Finnish and Northern Sámi as a case study to investigate how suitable multilingual LLMs are for low-resource machine translation and how much performance can be improved using supervised finetuning with varying amounts of parallel data. Our experiments on zero-shot translation reveal that mainstream multilingual LLMs from a variety of model families are unsuitable for translation between our chosen languages as-is, regardless of the generation hyperparameters. On the other hand, our experiments on supervised finetuning reveal that even relatively small amounts of parallel data can be very useful for improving performance in both translation directions.

How multilingual are multilingual LLMs? A case study in Northern Sámi-Finnish Translation

Emoji reactions are a frequently used feature of messaging platforms, yet their communicative role remains understudied. Prior work on emojis has focused predominantly on in-text usage, showing that emojis embedded in messages tend to amplify and mirror the author's affective tone. This evidence has often been extended to emoji reactions, treating them as indicators of emotional resonance or user sentiment. However, they may reflect broader social dynamics. Here, we investigate the communicative function of emoji reactions on Telegram. We analyze over 650k crypto-related messages that received at least one reaction, annotating each with sentiment, emotion, persuasion strategy, and speech act labels, and inferring the sentiment and emotion of emoji reactions using both lexicons and LLMs. We uncover a systematic mismatch between message and reaction sentiment, with positive reactions dominating even for neutral or negative content. This pattern persists across rhetorical strategies and emotional tones, indicating that emojis used as reactions do not reliably function as indicators of emotional mirroring or resonance of the content, in contrast to findings reported for in-text emojis. Finally, we identify the features that most predict emoji engagement. Overall, our findings caution against treating emoji reactions as sentiment labels, highlighting the need for more nuanced approaches in sentiment and engagement analysis.

Emoji Reactions on Telegram: Unreliable Indicators of Emotional Resonance

Large language models (LLMs) are now widely used in applications that depend on closed-ended decisions, including automated surveys, policy screening, and decision-support tools. In such contexts, these models are typically expected to produce consistent binary or ternary responses (for example, Yes, No, or Neither) when presented with questions that are semantically equivalent. However recent studies shows that LLM outputs can be influenced by relatively minor changes in prompt wording, raising concerns about the reliability of their decisions under paraphrasing. In this paper, we conduct a systematic analysis of paraphrase robustness across five widely used LLMs. To support this evaluation, we develop a controlled dataset consisting of 200 opinion-based questions drawn from multiple domains, each accompanied by five human-validated paraphrases. All models are evaluated under deterministic inference settings and constrained to a fixed Yes/No/Neither response format. We assess model behavior using a set of complementary metrics that capture the stability of each evaluated model. DeepSeek Reasoner and Gemini 2.0 Flash show the highest stability when responding to paraphrased inputs, whereas Claude 3.7 Sonnet exhibits strong internal consistency but produces judgments that differ more frequently from those of other models. By contrast, GPT-3.5 Turbo and LLaMA 3 70B display greater sensitivity to surface-level variations in prompt phrasing. Overall, these findings suggest that robustness to paraphrasing is driven more by alignment strategies and reasoning design choices than by model size alone.

Measuring LLMs’ Sensitivity to Paraphrased Opinion Prompts

Emotional tone plays a central role in persuasion, yet its impact on computational assessments of political argument quality in real world election campaign speeches remains understudied. In this work, we investigate whether positive emotional framing correlates with higher perceived convincingness in political arguments. We fine-tune language models on argument quality datasets and test their ability to transfer convincingness predictions to real-world campaign speeches. Using a corpus of U.S. presidential campaign speeches, we analyze emotional polarity in relation to predicted persuasive strength to test whether positively framed arguments are judged more convincing than neutral or negative ones. Our empirical analysis shows that political parties rely heavily on argumentation during their election campaigns. Also, we found the evidence that politicians strategically employ emotional cues within their arguments during these campaign speeches, with positive emotions being more strongly associated with persuasive strength, for example in topics such as USMCA’s Effect on American Jobs and Agriculture, Border Control Policies, Progressive Tax Reforms. At the same time, we find that negative emotions have a weaker yet still non-negligible influence on voter persuasion in topics such as City Crime and Civil Unrest and White Supremacist Violence (Charlottesville Incident).

Predicting Convincingness in Political Speech: How Emotional Tone Shapes Persuasive Strength

This study examines the capability of LLMs to predict emotional ratings of Russian words by comparing their assessments with both native speakers' ratings and expert evaluations. The research utilises two datasets: the ENRuN database containing associative emotional ratings of Russian nouns by native speakers, and RusEmoLex, an expert-compiled lexicon. Various open-source LLMs were evaluated, including international models (Llama-3, Qwen 2.5), Russian-developed models, and Russian-adapted variants, representing three parameter scales. The findings reveal distinct patterns in model performance: Russian-adapted models demonstrated superior alignment with native speakers' ratings, whilst model size was not a decisive factor. Conversely, larger models showed better performance in matching expert assessments, with language adaptation having minimal impact. Emotional or sensitive lexis with strong connotations produce a more substantial human-model gap.

Emotional Lexicons: How Large Language Models Predict Emotional Ratings of Russian Words

This submission describes a system developed for the AMIYA shared task targeting Syrian Arabic dialect modeling. The approach is based on parameter-efficient fine-tuning using Low-Rank Adaptation (LoRA) applied to a pretrained large language model, combined with prompt-guided inference to encourage fluent and natural dialectal output. The system is designed to prioritize colloquial Syrian Arabic and dialectal fidelity rather than strict factual accuracy, in alignment with the AL-QASIDA evaluation framework. Experimental outputs demonstrate improved fluency and reduced reliance on Modern Standard Arabic compared to the baseline model.

SDNLP at AMIYA 2026: Syrian Arabic Dialect Modeling with LoRA

In this paper, we describe models developed by our team, NUS-IDS, for the Closed data track at the Arabic Modeling In Your Accent (AMIYA) shared task at VarDial 2026. The core idea behind our solution involves data augmentation enabled by a dialect classifier trained on AMIYA data. We effectively combine various translation, summarization, and question answering prompts with AMIYA training data to form dialectal prompts for use with state-of-the-art LLMs. Next, dialect predictions from our classifier on outputs from these LLMs are used to compile preference data for Reinforcement Learning (RL). We report model performance on dialectal Arabic from Egypt, Morocco, Palestine, Saudi Arabia and Syria using FLORES+, a multilingual machine translation dataset. Our experiments illustrate that though our RL models show significant performance gains on dialectness scores, they under perform on translation metrics such as chrF++ compared to base LLMs.

NUS-IDS at AMIYA/VarDial 2026: Improving Arabic Dialectness in LLMs with Reinforcement Learning

We describe an open track system for modeling Palestinian Arabic that is developed for the AMIYA shared task using a parameter efficient fine tuning strategy. A 1.5B instruction-tuned language model was adapted with LoRA, updating only .28% of the model parameters, and trained on an aggregated set of conversations between Palestinians and resources covering both translation and generation. Model selection was guided by a comparative bench mark that prioritized performance efficiency and its tradeoffs. At the same time the paper focuses on targeting error analysis as well as structured instruction following. These findings illustrate both the viability and shed light on the current limitations of efficient adaptation methods for low resources Arabic dialects.

Premium content

Downloads

Next from EACL 2026 Main Conference

Grammatical Error Correction for Low-Resource Languages: The Case of Zarma

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES