How a student can choose a topic?

[pr-elhajji]

How user can chose what (s)he want to do with open tutor in a new topic (chat):

• TOPIC
• LEVEL
• Learnign objectives,
  ....

[Dakir-Ai]

If I understand the question correctly: The Open WebUI allows admins (or teachers) to create and personalize models. In our case, we can adapt a model for each topic, including specific knowledge and instructions for the model to follow. Users can easily switch between these (topics) as needed. By integrating the RLHF approach, the AI-model will continuously improve its ability to generate coherent and efficient content.

[TarekAi]

Thanks for your comment @dak-Aic 😊 Just to clarify, in OpenTutorAI, students will choose their topic and level (like a specific course or skill ..), and the model will generate personalized content based on those choices.
How do you think we can best structure the flow for topic and level selection, and what would the interface look like to make this process smooth and intuitive 🤔💡?

[Dakir-Ai]

@TarekAi Got it!

[Dakir-Ai]

We can streamline the topic and level selection process by integrating Open WebUI’s filters feature. This would allow users to customize their learning journey dynamically.

How Filters Can Be Used in OpenTutorAI ?

The filters feature in Open WebUI allows us to define custom filtering rules for our models. In our case, we want to filter available topics, levels, and languages to create a personalized learning experience.

✅ RLHF Integration: By incorporating Reward Models, the AI can learn from user interactions to improve content quality over time.

💡 UI Design : Step-by-step selection process (like a guided form).

[TarekAi]

Thanks for your suggestion @dak-Aic , The filtering feature is definitely a good tool, especially for handling incoming and outgoing messages. That said, for topic and level selection, filters might not be the best fit since they are more about modifying or monitoring interactions rather than shaping user choices.
I really like the step-by-step selection process. It keeps things clear and helps users navigate smoothly. Instead of filters, we could explore a more dynamic approach where topics and levels adapt based on user inputs. For example, the moment a user selects New Course or New Revision, we could instantly suggest relevant topics and levels instead of having them search manually.
What do you think ? Should we keep the structured step-by-step flow as it is, or do you see opportunities to make it more interactive 🤔 ?

[Dakir-Ai]

I came across this tool—Crossing Minds' RAGsys for LLM fine-tuning. Could we leverage it to solve our personalization challenges 🤔?

[TarekAi]

Using similar tools is a good idea, is it free ?

[Dakir-Ai]

No is not a free service. It's a commercial AI solution designed for businesses and educational institutions.

[TarekAi]

After a reflection on how to improve the learning experience using Open TutorAI, an idea comes in mind for both students and teachers ...
I think the best scenario is where the teacher sets up the topics, learning goals, and assignments, and students just show up and start learning.

1. Teachers can create classrooms, where tehy can add topics, set learning objectives, and give assignments based on the subjects they teach .. Teachers can also add AI tools to the classroom
2. Students can simply join the classroom and get started. No confusion, just direct access to the lessons and tasks they need to focus on

Each classroom will have topics set by the teacher. These topics will follow a clear, structured path to make sure students are always learning what they need to learn. Additionally, in the classroom, teachers can add AI tools like personal assistants (LLM) that help you with specific topics

Each topic will have clear learning objectives so students know exactly what they’re aiming for.

What do you think ?? 🤔

[XredaX]

This method for choosing a topic looks great! In my opinion, adding a short reflection exercise or quiz to help students identify their personal interests and goals could make the selection process even more personalized

[Dakir-Ai]

I like this vision @TarekAi , it keeps things structured for teachers while making learning straightforward for students. I hadn’t considered this approach before, as I’ve been focused on the technical workflow. That said, I agree with Reda’s suggestion to add short reflection exercises or quizzes to gather individual student data (Note: We prioritize robust safeguards to ensure complete data privacy and security). This could help personalize the learning experience further.