New technologies provide new ways to learn and practice languages

Background

Learning second languages is challenging. It can be difficult to find sufficient opportunities to practice. Even in formal learning environments, there is a limit to the amount of practice time that can be found and it is difficult to simulate conversations. Added to this, students are often self-conscious about their pronunciation or about making mistakes in front of their teachers or other students..
New natural language processing technologies offer a potential solution, as they can be readily accessible for anyone online and they are non-judgmental.

The Brief

The objective was to explore the opportunities available from new natural language processing technologies and develop an app that takes advantage of the ready availability of new technologies to develop an app that makes language learning more effective, while also being enjoyable to use.

Solution

The team worked to understand how to use newly available technologies to design functionality and a user experience that can significantly improve student experience and progression in attainment of fluency.

They developed the app for use on mobile and desktop that allows teachers to create lessons in any of six initial languages for their students to learn, practice and demonstrate their understanding and speaking abilities. Students can practice pronunciation, comprehension, translation and conversation. The app has the ability to provide structured role plays and also to have more unstructured interactions that can be created by teachers to mimic real world conversations, different to any other language learning technology on the market.

Technology

Building on Node.js and related services for the development of a highly usable, secure and scalable application, the app has Natural Language Processing technologies as its differentiator.

Amazon Polly is used to quickly and accurately convert text to natural sounding speech in a choice of supported voices. The Web Speech API is used for automatic speech recognition to convert speech to text and for natural language understanding. These services combine to give students natural language experiences, allowing users to learn and practice conversational components in both structured roleplays and unstructured chatbots.

A conversational editor built in React enables teachers to build structured roleplays and Rivescript is used for teachers to build chatbots for more natural unstructured conversation practice to bring their students capabilities to a new level.

Teachers gain insights into the engagement by their students with oral exercises, so they have an enhanced understanding of the progress that students are making and of any gaps in competence that they can bridge through more tailored lessons in the app.