HYDERABAD: Researchers at the Raj Reddy Centre for Technology and Society (RCTS), IIIT-H have designed a unique web-based tool to assist students and teachers in learning the English language, with an aim to make the language learning accessible to all.
The English language learning application is a machine learning, AI-based web application, which has been developed to assist classroom learning as well as self-learning of the language through speech interaction.
This machine learning model compares the speech of the expert and the speaker and gives an output with correct pronunciation.
Arjun Rajasekar, senior research scientist, RCTS, speaking to TNIE, said, “The tool can be used in a classroom setting as well as in self-learning exercise. In the classrooms, teachers can assess the performance of students with the help of the tool. In the self-learning mode, a student can use the expert speech which can be played repeatedly to learn or improve their language. In terms of application, when a student reads a sentence, the expert voice visualises the sentence and then gives an output with correct pronunciation for a mispronounced word with the expert voice, and gives feedback in the form of a score, which helps the students understand and improve the pronunciation of a particular word”.
The tool can save time and effort for teachers as they would not need repeat a word to correct the students. Where at times, the teachers themselves might not be proficient with the language and might find it difficult to explain the pronunciation, the tool can assist both teachers and students in improving their language skills.
The tool can work on any device that has a browser, microphone, speakers and an internet connection — either a desktop or a mobile phone.
Researchers have decided to use a speech assessment model developed by Prof Chiranjeevi Yarra from the Speech Processing Lab at the Language Technology Research Centre (LTRC) in IIITH.
The original model has the functionality of providing pronunciation feedback at four different levels — the phoneme, emphasis, tonal, and possible pause locations of a sentence, while the current implementation focuses only on the syllable stress component.
A pilot model test was conducted at a government high school in Moinabad in Rangareddy district in March, in which students in the sixth grade participated.
“We want to make English language learning reachable to a large number of audiences, especially those in rural areas. As of now, the tool is in pilot mode and we will be conducting more test runs in the coming times and based on feedback from teachers, modify the features and functionalities. Currently, the tool can record one sentence at a time. We are working to make it into long paragraphs and story-telling texts in the new modifications. The tool will eventually be made available in the public domain,” Rajasekhar added.
The tool as of now focuses on high school students from Classes 6 to 10, but scientists are also planning to make it available for primary classes in the future.