Background:
Reference: ALL v0.2 Release.
Subsequent to the release of v0.2 of Assisted Language Learning, the adoption user base increased to thousands of children. Feedback and improvement suggestions started to roll in. It is continued to focus on finely tuning and shaping up the simple and core mechanics of "speak with me".
Release objective:
- To fix bugs, improve user experiences, incorporate adoption feedback
- Establish foundational framework for personalised and adaptive learning
- Enhance telemetry capabilities
Persona
- Student/Learner who consumes ALL
- An administrator/coordinator who facilitates the usage of ALL for the Students. The coordinator could be someone at the School level, district level or at a State level
The Story
With increased adoption of the v0.2, thousands of Kayalvizhis started to engage with ALL and the data started flowing. Patterns started to emerge on the average duration of a session by the student, number of interactions, responses pattern in the application. With both quantitative and qualitative feedback and insights on the deployed version of ALL. It was evident that students were enjoying the simple, clean and interactive mechanics that's powered by the AI, and discover their capabilities and strive to continuously improve on their language learning abilities.
Use Cases
- System captures the Telemetry, both on the Adopter side (as a Local telemetry) and continue to publish anonymized telemetry into ALL servers (as Remote telemetry)
- System interjects to filter and mask any profanity in the language, either introduced by the user or system generated - in both Tamil and English
- Student has an improved user experience
- Student can login and have their individualized learning progress being tracked and captured
- Students are enabled with reward points and stars, based on the learning progress
- Students could express their emotion at the every RESPONSE - a like or a dislike
- Coordinator can monitor and analyse the RESPONSEs by the Students, from ALL v0.5
Key features:
- Upgraded to latest ASR by AI4Bharat -
- Profanity Service in Tamil and English
- Configurable Telemetry end points - on both ALL and Adopter sides
- Intuitive usage of Speak button
- User login
- User specific telemetry and learning progress
- User RESPONSE analytics on Words/Sentences
Tech Stack:
- A React JS based client side component
- Sunbird HTML/JS Telemetry specification integrated in React JS client code
- Node.js server endpoint
- Clickbase/MySQL database
- Metabase Visualization and Reporting
Released On:
29th September, 2023
Deployment/Hosting Environment:
- ALL v0.5 Hosted as an Appliance in the ALL Server - hosted by Ekstep Foundation
- ALL v0.5 Telemetry is hosted in the ALL Learning AI Server - hosted by Ekstep Foundation
