Sinau

Generative learning platform to help students learn the unthinkable. Name anything, we'll generate a multimodal content learning path to understand the impossible.

Comment

Inspiration

As educators and tech enthusiasts within Indonesia, we recognized the challenges faced by students in a diverse and vast country like ours. We observed that traditional education often struggled to cater to the unique needs and interests of individual learners, leading to disengagement and limited access to diverse knowledge. Inspired by the rich cultural heritage of Indonesia, we sought to create a platform that celebrates the nation's diversity and promotes inclusivity.

The name "Sinau," derived from the Javanese word meaning "to learn," embodies our commitment to honor Indonesia's linguistic and cultural richness. Our vision was to provide students across the archipelago with a personalized and limitless learning experience. We hoped that Sinau would ignite a passion for learning in every Indonesian student, inspiring them to unlock their full potential and contribute to the nation's progress with knowledge and confidence. By bridging the gap between traditional education and innovative technology, Sinau aspired to be a catalyst for positive change in Indonesia's education system, nurturing a generation of curious, creative, and well-rounded learners who are ready to face the future with an open mind and a thirst for knowledge.

What it does

  • We take any input on what topic you want to learn
  • We breakdown the given topic into subtopics using our own prompt utilizing OpenAI's GPT-4 model
  • We look and find the most relevant Youtube video for each generated subtopic using Youtube's Data v3 API
  • We transcribe the video to generate a quiz using OpenAI's Whisper model
  • We also find and summarize subtopic material using Duckduckgo's API to provide more reading material

How we built it

To create Sinau, we utilized a combination of technologies and APIs, the development process involved the following key steps:

  • Topic Breakdown: We used OpenAI's GPT-4 model to break down the user-provided topic into relevant subtopics. GPT-4 advanced language capabilities allowed us to understand the user's input better and generate coherent subtopics for a comprehensive learning experience.
  • Finding Relevant Youtube Video: Leveraging Youtube's Data v3 API, we searched for the most relevant videos for each generated subtopic. This ensured that learners could access multimedia content to complement their learning journey.
  • Video Transcription: OpenAI's Whisper model came in hand for transcribing the selected videos. The transcribed content was later used to create interactive quizzes, enhancing the learning process.
  • Summarizing Subtopic Material: To provide additional reading material, we utilized Duckduckgo's API to summarize information related to each subtopic. This allowed learners to dive deeper into the subject matter and gain a broader understanding.

Challenges we ran into

  • Deciding which basic features to MVP
  • Integrating each API

Accomplishments that we're proud of

Despite the challenges, we achieved several significant accomplishments throughout the development process:

  • Empowering curiosity and creativity, by providing personalized learning pathways, we inspired students to explore their interests, fostering curiosity and creativity in the learning process.

What we learned

The Power of AI in Education: AI and natural language processing can significantly enhance educational tools, making learning more accessible, personalized, and enjoyable for students. Adaptability and Collaborations: Flexibility and collaboration are essential when integrating multiple technologies.

What's next for Sinau

  • Fine-tuning prompt: We will continuously refine the prompt for the GPT-4 model to improve the accuracy and relevance of the generated subtopics.
  • Deploy our own custom-trained LLM model: Developing our own Language Model will give us greater control over customization and further optimize the learning experience.
  • Generative video using Synthesia: Introducing generative video content using Synthesia will enable us to create informative and engaging videos. Expanding Content and Reach: We aim to expand the platform's content library to cover a wider range of subjects and topics. Additionally, we plan to reach more students across Indonesia, including those in remote areas, to make quality education accessible to all.
Share this project:

Updates