3 research outputs found
A few-shot learning approach for a multilingual agro-information question answering system
No full textMini Dissertation (MIT (Big Data Science))--University of Pretoria, 2023.Agriculture plays a crucial role in numerous households across Sub-Saharan Africa.
Developing a question answering system that utilizes agricultural expertise and agroinformation can effectively bridge the support gap for farmers in the local community.
Most advances in question answering research involve large language models trained on
extensive data. Nevertheless, the conventional approach of fine-tuning has demonstrated
a significant decline in performance when models are fine-tuned on a small amount of
data. This decline is primarily attributed to the disparities between the objectives of pretraining and fine-tuning. One proposed alternative is to utilize prompt-based fine-tuning,
which permits the model to be fine-tuned with only a few examples. Extensive research
has been done on the application of these methods to tasks such as text classification
and not question answering. This research aims to study the feasibility of recent fewshot learning approaches, such as FewshotQA and Null prompting, for domain-specific
agricultural data in 4 South African languages. We evaluated the overall performance
of these approaches and investigated the effects of adapting these approaches for crosslingual extractive question answering of domain-specific data. The results obtained in
this study have shown valuable insight into the applicability of these methods to domainspecific data. These results have shown that these methods are capable of adequately
capturing the textual information of domain-specific data from the initial subset of data
points. Thus, there is potential for using these methods as a practical solution for limited
data.Computer ScienceMIT (Big Data Science)UnrestrictedFaculty of Engineering, Built Environment and Information Technolog
A Few‐Shot Learning Approach for a Multilingual Agro‐Information Question Answering System
No full textABSTRACT Across numerous households in Sub‐Saharan Africa, agriculture plays a crucial role. One solution that can effectively bridge the support gap for farmers in the local community is a question–answer system based on agricultural expertise and agro‐information. The more recent advancements in question answering research involve the use of large language models that are trained on an extensive amount of data. Due to this, conventional fine‐tuning approaches have demonstrated a significant decline in performance when using a significantly smaller amount of data. One proposed alternative to address this decline is to use prompt‐based fine‐tuning, which allows the model to be fine‐tuned with only a few examples, thus addressing the disparities between the objectives of pretraining and fine‐tuning. Extensive research has been done on these methods, specifically on text classification and not question answering. In this research, our objective was to study the feasibility of recent few‐shot learning approaches such as FewshotQA and Null‐prompting for domain‐specific agricultural data in four South African languages. We first explored creating a cross‐lingual domain‐specific extractive question answering dataset through an automated approach using the GPT model. Through exploratory data analysis, the GPT model was able to create a dataset, which requires minor improvements. We then evaluated the overall performance of the different approaches and investigated the effects of adapting these approaches to suit the new dataset. Results show these methods effectively capture semantic relationships and domain‐specific terminology but exhibit limitations, including potential biases in automated annotation and plateauing F1 scores. This highlights the need for hybrid approaches that combine artificial intelligence and human supervision. Beyond academic insights, this study has practical significance for industry, demonstrating how prompt‐based methods can help tailor AI models to specific use cases in low‐resource settings
Mafoko: Structuring and Building Open Multilingual Terminologies for South African NLP
No full textThe critical lack of structured terminological data for South Africa’s official languages hampers progress in multilingual NLP, despite the existence of numerous government and academic terminology lists. These valuable assets remain fragmented and locked in non-machine-readable formats, rendering them unusable for computational research and development. Mafoko addresses this challenge by systematically aggregating, cleaning, and standardising these scattered resources into open, interoperable datasets. We introduce the foundational Mafoko dataset, released under the equitable, Africa-centered NOODL framework. To demonstrate its immediate utility, we integrate the terminology into a Retrieval-Augmented Generation (RAG) pipeline. Experiments show substantial improvements in the accuracy and domain-specific consistency of English-to-Tshivenda machine translation for large language models. Mafoko provides a scalable foundation for developing robust and equitable NLP technologies, ensuring South Africa’s rich linguistic diversity is represented in the digital age.
