Morocco

We present SALT-31 benchmark dataset, for evaluation of machine translation (MT) models and covering 31 Ugandan languages. Unlike sentence-level evaluation sets, SALT-31 is constructed from short, scenario-driven mini-dialogues designed to preserve discourse context, pragmatics, and culturally grounded communication patterns common in everyday Ugandan settings. The dataset contains 100 English
sentences organized into 20 typical communication scenarios, each represented as a five sentence mini-sequence. It can therefore be used to evaluate both sentence-level and paragraph level machine translation, and includes nearly every language spoken in a country with high linguistic diversity.

EACL 2026

SALT-31: A Machine Translation Benchmark Dataset for 31 Ugandan Languages

ugandan languages

multilingual evaluation

dialogue context

benchmark dataset

low-resource languages

machine translation

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. 

This paper introduces the first standardized benchmark for evaluating Automatic Speech Recognition (ASR) in the Bambara language, utilizing one hour of professionally recorded Malian constitutional text. Designed as a controlled reference set under near-optimal acoustic and linguistic conditions, the benchmark was used to evaluate 37 models, ranging from Bambara-trained systems to large-scale commercial models. Our findings reveal that current ASR performance remains significantly below deployment standards; the top-performing system in terms of Word Error Rate (WER) achieved 46.76\% and the best Character Error Rate (CER) of 13.00\% was set by another model, while several prominent multilingual models exceeded 100\% WER due to severe hallucinations. These results suggest that multilingual pre-training and model scaling alone are insufficient for underrepresented languages. Furthermore, because this dataset represents a best-case scenario of the most simplified and formal form of spoken Bambara, these figures likely establish an upper bound for performance in practical, real-world settings. We provide the benchmark and an accompanying public leaderboard to facilitate transparent evaluation and future research in Bambara speech technology.

Where Are We at with Automatic Speech Recognition for the Bambara Language?

Multimodal AI research has overwhelmingly focused on high-resource languages, hindering the democratization of advancements in the field. To address this, we present AfriCaption, a comprehensive framework for multilingual image captioning in 20 African languages and our contributions are threefold: (i) a curated dataset built on Flickr8k, featuring semantically aligned captions generated via a context-aware selection and translation process; (ii) a dynamic, context-preserving pipeline that ensures ongoing quality through model ensembling and adaptive substitution; and (iii) the AfriCaption model, a 0.5B parameter
vision-to-text architecture that integrates SigLIP and NLLB200 for caption generation across underrepresented languages. This unified framework ensures ongoing data quality and establishes the first scalable image-captioning resource for underrepresented African languages, laying the groundwork for truly inclusive multimodal AI.

AfriCaption: Establishing a New Paradigm for Image Captioning in African Languages

Existing AI bias evaluation benchmarks largely reflect Western perspectives, leaving African contexts underrepresented and enabling harmful stereotypes in applications across various domains. To address this gap, we introduce AfriStereo, the first open-source African stereotype dataset and evaluation framework grounded in local socio-cultural contexts. Through community engaged efforts across Senegal, Kenya, and Nigeria, we collected 1,163 stereotypes spanning gender, ethnicity, religion, age, and profession. Using few-shot prompting with human-in-the-loop validation, we augmented the dataset to over 5,000 stereotype–antistereotype pairs. Entries were validated through semantic clustering and manual annotation by culturally informed reviewers. Preliminary evaluation of language models reveals that nine of eleven models exhibit statistically significant bias, with Bias Preference Ratios (BPR) ranging from 0.63 to 0.78 (p ≤ 0.05), indicating systematic preferences for stereotypes over antistereotypes, particularly across age, profession, and gender dimensions. Domain-specific models appeared to show weaker bias in our setup, suggesting task-specific training may mitigate some associations. Looking ahead, AfriStereo opens pathways for future research on culturally grounded bias evaluation and mitigation, offering key methodologies for the AI community on building more equitable, context-aware, and globally inclusive NLP technologies.

AfriStereo: A Culturally Grounded Dataset for Evaluating Stereotypical Bias in Large Language Models

Language choice in multilingual societies is rarely arbitrary. In Nigerian, English, Nigerian Pidgin (NP) and indigenous languages are strategically deployed in online discourse, yet little is known about how they function in hostile contexts. Here we conduct the first systematic analysis of NP in online hate speech on two platforms, Twitter and Instagram. Using a linguistically enriched annotation scheme, we label each post for class, targeted group, language variety, and hate type. Our results show that NP is disproportionately used in offensive and hateful discourse, particularly against Hausa, women, and LGBTQ+ groups, and that insults are the dominant hate strategy. Cross-domain evaluation further reveals that classifiers trained on Twitter systematically over-predict hate on Instagram, highlighting challenges of domain transfer. These findings underscore NP’s role as a linguistic resource for hostility and its sociolinguistic salience in amplifying stereotypes and affect. For NLP, the work demonstrates the need for NP-specific resources, sensitivity to figurative strategies, and domain adaptation across platforms. By bridging sociolinguistics and computational modeling, this study contributes new evidence on how language choice shapes online hate speech in a multilingual African context.

Language Choice in Nigerian Social Media Hate Speech

Low-resourced languages are increasingly included in large multilingual models. While including more languages in pretrained models is a sign of progress, large models still underperform on low-resourced languages. In prioritizing scale over effective processing, we risk 1) deploying language technologies that misrepresent these languages and 2) amplifying gender biases embedded in training corpora. In this paper, we investigate how masked language models encode gender for three low-web-data languages, Afan Oromo, Amharic, and Tigrinya, and how these representations shift after continued pretraining on NLLB data. Using a controlled cloze-style probing setup, we examine prediction patterns. Our findings show consistent gender asymmetries and predictions aligned with stereotypical adjectives and occupations. After continued pretraining, we find that male-gendered predictions reach up to 68% in Amharic, while neutral predictions exceed 60% in Afan Oromo. Our work shows that expanding training data does not guarantee balanced gender representations without careful consideration in data curation.

Probing Gender Bias in Masked Language Models for Low-Web Data Languages

We present Kunkado, a 160-hour Bambara ASR dataset compiled from Malian radio archives to capture present-day spontaneous speech across a wide range of topics. It includes code-switching, disfluencies, background noise, and overlapping speakers that practical ASR systems encounter in real-world use. We finetuned Parakeet-based models on a 33.47-hour human-reviewed subset and apply pragmatic transcript normalization to reduce variability in number formatting, tags, and code-switching annotations. Evaluated on two real-world test sets, finetuning with Kunkado reduces WER from 44.47\% to 37.12\% on one and from 36.07\% to 32.33\% on the other. In human evaluation, the resulting model also outperforms a comparable system with the same architecture trained on 98 hours of cleaner, less realistic speech. We release the data and models to support robust ASR for predominantly oral languages.

Kunnafonidilaw ka Cadeau: an ASR dataset of present-day Bambara

Most of the more than 40 languages spoken in Togo lack the data and tools necessary for modern natural language processing (NLP) applications. We present an extension of previous work by introducing new datasets, improved models, and a community-driven evaluation benchmark for text-to-speech (TTS). We expanded the Eyaa-Tom multilingual corpus with additional speech data (e.g. 26.9k recordings, 30.9 hours) across 10 local languages and incorporated Mozilla Common Voice contributions (64.6k clips, 46.6 hours) for Adja, Nawdm, Mina, Tem to strengthen automatic speech recognition (ASR) and speech synthesis. We detail how community contributors (including collaboration with a national TV journalist) helped collect and validate the Kabiyɛ and French text, with an ethical compensation model in place. We also try to compare the performance of a few models in these datasets, we fine-tuned state-of-the-art models in these data for ASR, OpenAI Whisper and faster-whisper were benchmarked achieving improved word error rates after fine-tuning; for machine translation, we fine-tuned Meta's NLLB-200 model in 11 local languages, which produced significant BLEU/METEOR gains especially in Ewɛ, and Kabɩyɛ. To evaluate TTS, we introduce Lom Bench, a new community-based benchmark where native speakers rate synthetic speech. The preliminary results from Lom Bench indicate promising naturalness in Ewɛ and Kabɩyɛ TTS, although further data is needed.

Eyaa-Tom 26, Yodi-Mantissa and Lom Bench: A Community Benchmark for TTS in Local Languages

In this work, we present AfriNLLB, a series of lightweight models for efficient translation from and into African languages. AfriNLLB supports 15 language pairs (30 translation directions), including Swahili, Hausa, Yoruba, Amharic, Somali, Zulu, Lingala, Afrikaans, Wolof, and Egyptian Arabic, as well as other African Union official languages such as Arabic (MSA), French, Portuguese, and Spanish. Our training data covers bidirectional translation between English and 13 languages, and between French and two languages (Lingala and Wolof). 

AfriNLLB models are based on NLLB-200. We compress the NLLB-200 600M with two approaches, iterative layer pruning and quantization. Our work aims at enabling efficient deployment of translation models for African languages in resource-constrained settings. 
Our evaluation results demonstrate that AfriNLLB models achieve performance comparable to the baseline while being over 20\% times faster. We release two versions of the AfriNLLB models, a Transformers version that allows further fine-tuning and a CTranslate2 version for efficient inference. Moreover, we release all the training data that we curated for fine-tuning the baseline and pruned models to facilitate further research.

AfriNLLB: Efficient Translation Models for African Languages

Large language models are extremely sensitive to prompt design, a phenomenon which is amplified in multilingual scenarios especially with low-resource languages due to low coverage in model training data, orthographic variation and tokenization issues.
In this work, we evaluate a reflective prompt evaluation technique (GEPA) as an inference time optimization strategy on multiple multilingual benchmarks spanning various African languages.
Using a more capable model as a reflection model and operating under strict optimization budgets - we show that reflective prompt optimization enables inference time improvements through textual policy evolution resulting in consistent improvements on different tasks without any weight updates. 

Our evaluation mostly focuses on closed-source models and we observe that on some of the benchmarks, prompt-optimized smaller models can outperform better and more recent models highlighting the importance of instruction design for measuring the capabilities of a language model on a given task.
Qualitative review of model outputs before and after optimization also shows that prompt evolution not only reinforces models ability to perform a particular task, it also improves output formatting which is very important for proper model evaluation.
We characterize the resulting prompt-latency tradeoff by quantifying the optimization cost in terms of prompt token growth, and our results show that modest increase in prompt size can result in substantial gains in performance. Finally, we argue that benchmark evaluations should report prompt-optimized results alongside baseline prompts in order to properly reflect model capabilities for low-resource languages.

What Do Prompts Reveal About Model Capabilities in Low-Resource Languages?

BambaraMLLM is a unified multilingual multimodal large language model (MMLLM) designed to address the critical lack of digital resources for Bambara, a West African language spoken by over 15 million people. Unlike traditional approaches that rely on task-specific models for different linguistic functions, BambaraMLLM integrates text generation, automatic speech recognition (ASR), machine translation (MT), and text-to-tpeech (TTS) synthesis into a single, transformer-based architecture. This work establishes a scalable, open-source foundation for African language technology, optimizing for both performance and deployment under resource constraints.

Downloads

Next from EACL 2026

Where Are We at with Automatic Speech Recognition for the Bambara Language?

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES