ReadSmart

Inspiration

The inspiration behind ReadSmart came from the need for an interactive and supportive tool that helps children build strong reading comprehension and pronunciation skills. Many young learners face difficulties with reading fluency and understanding, often lacking personalized feedback in traditional educational settings. Our goal was to leverage AI to make learning engaging, accessible, and tailored to each student's needs, encouraging a lifelong love for reading.

What it does

ReadSmart is an AI-powered web app designed to help children improve their reading and speaking skills:

• Customized Reading Passages: It generates age-appropriate reading passages based on a grade 5 curriculum, introducing themes like friendship, nature, and curiosity.
• Pronunciation Feedback: As children read aloud, the app listens, identifies mispronunciations, and provides gentle guidance for improvement.
• Interactive Q&A: After reading, the app engages students in comprehension questions tailored to the passage, reinforcing understanding through a conversational, supportive approach.

How we built it

• Language Model (LLM): We used OpenAI's GPT-4 for generating reading passages and comprehension questions dynamically, allowing fresh content each time.
• Pronunciation Assessment: Azure’s Pronunciation Assessment API analyzes phoneme-level accuracy and provides a detailed breakdown of areas for improvement.

Challenges we ran into

• Phoneme-Level Accuracy: Ensuring accurate detection of each phoneme in children’s voices proved challenging, especially when factoring in variations in accents and tone.
• Dynamic Content Generation: Designing prompts that consistently generate age-appropriate, engaging content required substantial fine-tuning and testing.
• Real-Time Feedback Timing: Balancing real-time responsiveness with detailed feedback, without disrupting the reading flow, was a delicate task.

Accomplishments that we're proud of

• Responsive Pronunciation Feedback: Our app’s ability to give precise, phoneme-level feedback is a significant achievement, allowing kids to improve specific sounds.
• Personalized Learning Path: By generating unique passages and questions each session, we’ve created a dynamic, engaging educational tool tailored to individual progress.

What we learned

• Effective Prompt Engineering: Crafting prompts to generate consistently age-appropriate and engaging material was key to making the app effective for its target audience.
• Complexities of Speech Analysis: We gained insight into the intricacies of phoneme analysis and the technical requirements for accurate pronunciation assessment in young voices.

What's next for ReadSmart

• Expanding to Multiple Grade Levels: We plan to extend content options to cover more grade levels, creating a comprehensive learning tool across elementary education.
• Support for Additional Languages: Adding multi-language capabilities will make ReadSmart a versatile tool for students around the world.
• Gamification: Introducing gamified elements, like achievements and progress tracking, will make learning more engaging and fun for young users.
• Parent & Teacher Dashboard: Providing a dashboard where parents and teachers can monitor student progress and customize learning goals will enhance the app’s utility in both home and classroom settings.

Built With

• azure
• openai