Machine Translation (MT) for low-resource languages, such as Twi, remains a persistent challenge in natural language processing (NLP) due to the scarcity of extensive parallel datasets. Due to their heavy reliance on high-resource data, traditional methods frequently fall short, underserving low-resource languages. To address this, we propose a fine-tuned T5 model trained using a Cross-Lingual Optimization Framework (CLOF), a unique method that dynamically modifies gradient weights to balance low-resource (Twi) and high-resource (English) datasets. This cross-lingual learning framework leverages the strengths of federated training to improve translation performance while ensuring scalability for other low-resource languages. In order to maximize model input, the study makes use of a carefully selected parallel English-Twi corpus that has been aligned and tokenized. A thorough evaluation of translation quality is provided by the use of SPBLEU, ROUGE (ROUGE-1, ROUGE-2, and ROUGE-L) measures, and Word Error Rate (WER) metrics. A pretrained mT5 model is used to set baseline performance, which acts as a standard for the optimized model. The suggested method shows notable benefits, according to experimental results. The fine-tuned model achieves a remarkable increase in SPBLEU from 2.16% to 71.30%, a rise in ROUGE-1 from 15.23% to 65.24%, and a notable reduction in WER from 183.16% to 68.32%. These findings highlight the effectiveness of CLOF in addressing the challenges of low-resource MT and enhancing the quality of Twi translations. This work demonstrates the potential of combining cross-lingual learning and federated training to advance NLP for underrepresented languages, paving the way for more inclusive and scalable translation systems.
Civil structural component detection is an important stage in structural health monitoring for the interpretation and analysis of damages associated with various structural components. Structural component detection is a challenging task due to the asymmetrical shape of the components. We introduce Gabor filters as an enhancement for a one-stage detector. Generic Gabor filters are generated to replace the lower convolutional filters of the backbone of SSD. Evaluation of the variant SSD detector on structural imagery shows promising detection results. The experimental results show an improvement in accuracy and how fast the variant SSD detector learns. A comparison with some state-of-the-art detectors indicates the SSD with Gabor filters is favorable.
12 August 2024 in IEEE Journal of Quantum Electronics
This research delves into the innovative application of the optical Stark effect in dynamically guiding polaritons through a Y-shaped potential, forming a polaritonic coupler within a 2D hybrid organic-inorganic perovskite microcavity. The study explores the characteristics of the 2D perovskite, focusing on harnessing the optical Stark-induced energy shift in the polariton branches. The polaritonic coupler, which has a single input and two divergent outputs, is subjected to an external optical Stark pulse, dynamically guiding polaritons between the input and outputs. The research focuses on examining the controllability of the polaritonic coupler through the polariton coupling ratio, highlighting the regulatory role played by the optical Stark effect in this dynamic process. In-depth analyses of the spatial distribution and time evolution of polaritons within the coupler reveal that the optical Stark pulse effectively regulates the polariton coupling ratio, realizing a programmable coupler. This investigation not only advances the fundamental understanding of polariton dynamics within 2D hybrid organic-inorganic perovskite microcavities but also demonstrates the potential for developing optically controlled integrated photonic devices.
4 October 2025 in Biomedical Signal Processing and Control
Machine Translation (MT) for low-resource languages, such as Twi, remains a persistent challenge in Natural Language Processing (NLP) due to the scarcity of extensive parallel datasets. Due to their heavy reliance on high-resource data, traditional methods frequently fall short, underserving low-resource languages. To address this, we propose a fine-tuned T5 model trained using Cross-Lingual Optimization Framework (CLOF), a unique method that dynamically modifies gradient weights to balance low-resource (Twi) and high-resource (English) datasets. This cross-lingual learning framework leverages the strengths of federated training to improve translation performance while ensuring scalability for other low-resource languages. In order to maximize model input, the study makes use of a carefully selected parallel English-Twi corpus that has been aligned and tokenized. A thorough evaluation of translation quality is provided by the use of SPBLEU, ROUGE (ROUGE-1, ROUGE-2, and ROUGE-L) measures, and Word Error Rate (WER) metrics. A pretrained mT5 model is used to set baseline performance, which acts as a standard for the optimized model. The suggested method shows notable benefits, according to experimental results. The fine-tuned model achieves a remarkable increase in SPBLEU from 2.16% to 71.30%, a rise in ROUGE-1 from 15.23% to 65.24%, and a notable reduction in WER from 183.16% to 68.32%. These findings highlight the effectiveness of CLOF in addressing the challenges of low-resource MT and enhancing the quality of Twi translations. This work demonstrates the potential of combining cross-lingual learning and federated training to advance NLP for underrepresented languages, paving the way for more inclusive and scalable translation systems.