English Speaking AI-examiner

• 

  UI

Inspiration

As a non-native English speaker who migrated to Canada, I've been on a mission to improve my English skills. I was motivated to enhance my English speaking skills, but the high costs of language courses discouraged me. That's when I got the idea to create an app that could help me improve without financial constraints.

What it does

This application aims to provide a comprehensive language learning experience, focusing specifically on improving English speaking skills for test takers, such as those preparing for the IELTS exam. The application utilizes the Whisper model to transcribe the user's spoken responses. To provide a comprehensive evaluation, the application leverages a Language Learning Model (LLM) as well. The LLM takes both the transcribed text and the corresponding question as input. By considering the context of the question and the user's response, the LLM generates an assessment of the given answer. This assessment includes analyzing grammar, vocabulary usage, coherence, and fluency, among other language aspects.

How I built it

With my newfound knowledge, I started developing the project. Using Google Colab, I integrated the "Whisper" model with a Language Learning Model (LLM) to provide feedback on English language usage. It was an exciting journey as I combined these technologies to create a useful tool for language learners.

Challenges I ran into

Creating the app wasn't easy. I faced several challenges along the way. Being new to these tools, I spent days reading and completing courses to understand LangChain and LLMs better. It was a steep learning curve, but my determination kept me going.

Accomplishments that I'm proud of

Although this first version of this idea is still very premature, I feel proud to be able to showcase a result. Having had to study such innovative tools as LangChain and the use of LLMs in such a short time makes me proud, but above all, it motivates me to continue studying this subject because I believe it has a great future.

What I learned

To embark on this endeavor, I delved into understanding the intricacies of the IELTS speaking test format and evaluation criteria. This involved studying the various sections, question types, and scoring systems employed in the test. Additionally, I expanded my knowledge of speech recognition technology, language processing algorithms, and user interface design to create an effective and user-friendly simulation tool.

What's next for English Speaking AI-examiner

As I aspire to participate in international exams like the IELTS, I recognize the importance of thorough preparation. The IELTS test, especially the speaking component, can be challenging and intimidating. It requires not only language proficiency but also confidence and familiarity with the test format. Inspired by this goal, I decided to create a tool that would simulate the speaking test process, providing free preparation resources and instilling the necessary confidence to face the actual exam.

Built With

• dolly
• gradio
• huggingface
• langchain
• llms
• openai
• python
• whisper