Sorsolingo: A Bicol dialect machine translator
Project Type:
Academic Project
Tools Used:
Figma
Year:
2023
Samples
Situation:
The project aimed to address the language barrier between speakers of the Bisakol dialect, native to Southern Sorsogon, and those who speak English or Filipino. There was a need for a tool that could facilitate communication and learning of the Bisakol dialect, which would help in promoting the local culture and language, as well as supporting tourists, government institutions, and other users interested in learning the dialect.
The task:
I collaborated with a team to developed Sorsolingo, a mobile application designed to provide translation services from Bisakol to English and Filipino, and vice versa. The app needed to incorporate features such as text-to-speech and speech-to-text functionalities, store user search history, and provide a word list feature for easier learning.
Implementing the project:
Our team compiled datasets of Bisakol grammar, words, and phrases from local literary texts and online sources. We also evaluated the accuracy of the translations using machine learning algorithms such as MarianMT. After training the model, we integrated key features like text-to-speech and speech-to-text, ensuring users could practice proper pronunciation and understand spoken inputs. We also designed a user-friendly interface to enhance accessibility, making it convenient for users to learn and communicate in the dialect.
Result:
The result was a functional and accessible mobile translation tool that provides both written and spoken translations between Bisakol, English, and Filipino. The project fostered cultural understanding, making it easier for people to communicate in the Bisakol dialect, and promoted the preservation of the local language. It successfully addressed the language barriers in the Sorsogon area and provided a valuable educational resource.