Search

Low-resource languages often face challenges in acquiring high-quality language data due to the reliance on translation-based methods, which can introduce the translationese effect. This phenomenon results in translated sentences that lack fluency and naturalness in the target language. In this paper, we propose a novel approach for data collection by leveraging storyboards to elicit more fluent and natural sentences. Our method involves presenting native speakers with visual stimuli in the form of storyboards and collecting their descriptions without direct exposure to the source text. We conducted a comprehensive evaluation comparing our storyboard-based approach with traditional text translation-based methods in terms of accuracy and fluency. Human annotators and quantitative metrics were used to assess translation quality. The results indicate a preference for text translation in terms of accuracy, while our method demonstrates worse accuracy but better fluency in the language focused., Comment: published at LREC-COLING 2024

This paper investigates the role of text categorization in streamlining stopword extraction in natural language processing (NLP), specifically focusing on nine African languages alongside French. By leveraging the MasakhaNEWS, African Stopwords Project, and MasakhaPOS datasets, our findings emphasize that text categorization effectively identifies domain-agnostic stopwords with over 80% detection success rate for most examined languages. Nevertheless, linguistic variances result in lower detection rates for certain languages. Interestingly, we find that while over 40% of stopwords are common across news categories, less than 15% are unique to a single category. Uncommon stopwords add depth to text but their classification as stopwords depends on context. Therefore combining statistical and linguistic approaches creates comprehensive stopword lists, highlighting the value of our hybrid method. This research enhances NLP for African languages and underscores the importance of text categorization in stopword extraction., Comment: A Project Report for the Masakhane Research Community

African languages have far less in-language content available digitally, making it challenging for question answering systems to satisfy the information needs of users. Cross-lingual open-retrieval question answering (XOR QA) systems -- those that retrieve answer content from other languages while serving people in their native language -- offer a means of filling this gap. To this end, we create AfriQA, the first cross-lingual QA dataset with a focus on African languages. AfriQA includes 12,000+ XOR QA examples across 10 African languages. While previous datasets have focused primarily on languages where cross-lingual QA augments coverage from the target language, AfriQA focuses on languages where cross-lingual answer content is the only high-coverage source of answer content. Because of this, we argue that African languages are one of the most important and realistic use cases for XOR QA. Our experiments demonstrate the poor performance of automatic translation and multilingual retrieval methods. Overall, AfriQA proves challenging for state-of-the-art QA models. We hope that the dataset enables the development of more equitable QA technology.

AfriSenti-SemEval Shared Task 12 of SemEval-2023. The task aims to perform monolingual sentiment classification (sub-task A) for 12 African languages, multilingual sentiment classification (sub-task B), and zero-shot sentiment classification (task C). For sub-task A, we conducted experiments using classical machine learning classifiers, Afro-centric language models, and language-specific models. For task B, we fine-tuned multilingual pre-trained language models that support many of the languages in the task. For task C, we used we make use of a parameter-efficient Adapter approach that leverages monolingual texts in the target language for effective zero-shot transfer. Our findings suggest that using pre-trained Afro-centric language models improves performance for low-resource African languages. We also ran experiments using adapters for zero-shot tasks, and the results suggest that we can obtain promising results by using adapters with a limited amount of resources., Comment: SemEval 2023

Recent advances in the pre-training of language models leverage large-scale datasets to create multilingual models. However, low-resource languages are mostly left out in these datasets. This is primarily because many widely spoken languages are not well represented on the web and therefore excluded from the large-scale crawls used to create datasets. Furthermore, downstream users of these models are restricted to the selection of languages originally chosen for pre-training. This work investigates how to optimally leverage existing pre-trained models to create low-resource translation systems for 16 African languages. We focus on two questions: 1) How can pre-trained models be used for languages not included in the initial pre-training? and 2) How can the resulting translation models effectively transfer to new domains? To answer these questions, we create a new African news corpus covering 16 languages, of which eight languages are not part of any existing evaluation dataset. We demonstrate that the most effective strategy for transferring both to additional languages and to additional domains is to fine-tune large pre-trained models on small quantities of high-quality translation data., Comment: Accepted to NAACL 2022 (added evaluation data for amh, kin, nya, sna, xho)

With the success of large-scale pre-training and multilingual modeling in Natural Language Processing (NLP), recent years have seen a proliferation of large, web-mined text datasets covering hundreds of languages. We manually audit the quality of 205 language-specific corpora released with five major public datasets (CCAligned, ParaCrawl, WikiMatrix, OSCAR, mC4). Lower-resource corpora have systematic issues: At least 15 corpora have no usable text, and a significant fraction contains less than 50% sentences of acceptable quality. In addition, many are mislabeled or use nonstandard/ambiguous language codes. We demonstrate that these issues are easy to detect even for non-proficient speakers, and supplement the human audit with automatic analyses. Finally, we recommend techniques to evaluate and improve multilingual corpora and discuss potential risks that come with low-quality data releases., Comment: Accepted at TACL; pre-MIT Press publication version

We take a step towards addressing the under-representation of the African continent in NLP research by creating the first large publicly available high-quality dataset for named entity recognition (NER) in ten African languages, bringing together a variety of stakeholders. We detail characteristics of the languages to help researchers understand the challenges that these languages pose for NER. We analyze our datasets and conduct an extensive empirical evaluation of state-of-the-art methods across both supervised and transfer learning settings. We release the data, code, and models in order to inspire future research on African NLP., Comment: Accepted to TACL 2021, pre-MIT Press publication version

Massively multilingual machine translation (MT) has shown impressive capabilities, including zero and few-shot translation between low-resource language pairs. However, these models are often evaluated on high-resource languages with the assumption that they generalize to low-resource ones. The difficulty of evaluating MT models on low-resource pairs is often due to lack of standardized evaluation datasets. In this paper, we present MENYO-20k, the first multi-domain parallel corpus with a special focus on clean orthography for Yor\`ub\'a--English with standardized train-test splits for benchmarking. We provide several neural MT benchmarks and compare them to the performance of popular pre-trained (massively multilingual) MT models both for the heterogeneous test set and its subdomains. Since these pre-trained models use huge amounts of data with uncertain quality, we also analyze the effect of diacritics, a major characteristic of Yor\`ub\'a, in the training data. We investigate how and when this training condition affects the final quality and intelligibility of a translation. Our models outperform massively multilingual models such as Google ($+8.7$ BLEU) and Facebook M2M ($+9.1$ BLEU) when translating to Yor\`ub\'a, setting a high quality benchmark for future research., Comment: Accepted to MT Summit 2021 (Research Track)

Each year, as the title of this report suggests, UNCF awards well over 10,000 scholarships, fellowships and internships under 400 programs. As UNCF receives no federal or state funds, these programs are supported by contributions from corporations, foundations, families and individuals-- investments in the futures of the students who receive the scholarships, fellowships and internships. And because the nation and every community in it need the college-educated professionals students will become upon graduation, these are investments in everyone's futures. This report, issued biennially by UNCF's Frederick D. Patterson Research Institute (FDPRI), is UNCF's accounting for these programs collectively and, for several of the largest and most innovative programs, individually as well. UNCF's scholarship programs support students at UNCF-member institutions and other colleges and universities during a critical period of their academic lives. These scholarships eliminate a significant portion of the students' unmet financial need and therefore are instrumental in their ability to remain enrolled in and graduate from their institutions.

An appendix provides regression analyses.As the cost of higher education continues to climb, colleges and universities are finding themselves under increased pressure to demonstrate their value by enrolling, graduating and launching graduates into careers with the skills needed to compete as part of the American workforce. Few institutions are positioned with all of the resources required to support student success absent the public funding (be it student- or institution-based) that has long subsidized the cost of earning a college degree. These difficulties could not be truer of the fiscal environment in which historically black colleges and universities (HBCUs) operate. For more than 150 years, HBCUs have played a significant role in providing access to postsecondary education for millions of blacks, many of whom currently come from low-income families and are first-generation college students. As pressures mount for all institutions of higher education to be more accountable for demonstrating favorable student outcomes (e.g., retention and graduation rates), HBCUs and other minority-serving institutions face unique challenges that have often been overlooked in evaluations of institutional performance in retaining and graduating students. In this report, the authors provide an empirically based case for why evaluations of institutional performance should account for factors--namely, student characteristics, given their strong association with student outcomes related to persistence and completion--that are traditionally excluded in most measures of institutional performance. An appendix provides details on regression analyses. [For "Understanding HBCU Retention and Completion: UNCF-Member Institution Supplement," see ED562058.]

Despite the importance of postsecondary education to the economic and social vitality of the U.S. and the individuals who pursue this academic goal, the educational pipeline to and through college is broken for communities of color, the fastest-growing segment of the population. This report offers a revealing glimpse of the American system of public education from the perspective of those with the biggest stake in better schools: the parents of the African American children who are least well served by the system. This report is a meta-analysis of a two-part research program that involved both a quantitative survey and qualitative focus groups. The study was conducted in February and March 2012 in five target cities: Atlanta, Detroit, Memphis, New Orleans, and Washington, D.C. These cities were chosen because they have large African American populations, robust education reform efforts are under way or are part of a planning process to address the cities' challenges, and United Negro College Fund (UNCF) has a strong institutional presence there. These cities in particular face many of the educational challenges that prevent too many African American children from fulfilling their potential, and they therefore serve as excellent testing grounds. The key findings from this study reinforce and add greater details to the body of knowledge about African American parents and their involvement in the education of their children. Most important, a substantial percentage of low-income African American parents and caregivers who participated in the study (87 percent) have high aspirations for their children and overwhelmingly want them to graduate from college. An appendix contains the following: (1) Data Collection Methods; and (2) Topics on Which Low-Income Parents Would Like More Information to Support Their Child's Academic Progress.