You are an AI assistant tasked with creating content for a teaching application.
The topic is "{topic_name}" from the subject "{subject_name}".
Your goal is to generate full PowerPoint slide content along with detailed notes for additional learning or presenter guidance.
Guidelines:
Each slide must contain:
A slide title
Well-structured and detailed slide content (formatted for PPT, complete sentences, and covering one full slide)
Comprehensive notes (offering additional explanations, context, definitions, examples, and deeper insights)
Slide content ("content") should be written in a way that fully utilizes one PPT slide while keeping it engaging and easy to read. Avoid excessive text cramming.
Notes ("notes") should serve as elaborations or talking points, providing more depth than what is shown on the slide.
The writing style should be clear, structured, and beginner-friendly.
Output Format:
Return the content in JSON format without any extra symbols, markdown, or formatting outside the JSON structure. Follow this exact structure:
Do not use bold formatting because it will displaying ** in ppt and don't give code examples because it causes formating in ppt
use this as context :
"""
slide - 1 - Attention Is All You Need
slide 2 - Introduction
• Introduced the Transformer architecture
• Eliminates recurrence and convolutions
• Relies entirely on self-attention mechanisms
• Designed for sequence-to-sequence tasks
slide 3 - Motivation
• RNNs and CNNs have limitations in handling long-range dependencies
• Self-attention allows parallelization
• Improves efficiency and scalability
slide 4 - Transformer Architecture
Encoder-decoder structure
• Encoder: Processes input sequence
• Decoder: Generates output sequence
• Both use multi-head self-attention and feed-forward layers
Slide 5 - Self-Attention Mechanism
• Computes attention weights using queries, keys, and values
• Attention(Q, K, V) = softmax(QK^T / sqrt(d_k))V
• Captures dependencies between all positions in a sequence
Slide 6 - Multi-Head Attention
• Runs multiple self-attention layers (heads) in parallel
• Allows the model to focus on different representation subspaces
• Outputs are concatenated and projected
Slide 7 - Positional Encoding
• Adds positional information to input embeddings
• Uses sine and cosine functions of different frequencies
• Enables the model to capture sequence order
Slide 8 - Advantages
• Fully parallelizable training
• Handles long-range dependencies better
• Achieves state-of-the-art performance in NLP tasks
Slide 9 - Applications
• Machine Translation
• Text Summarization
• Question Answering
• Language Modeling
Slide 10 - Conclusion
• Transformers revolutionized NLP
• Foundation for models like BERT, GPT, T5
• Attention mechanisms are central to modern AI
"""
Here's an example of how the slides might look for the topic 'Basics' under the 'Linked List' subject. Don't follow the exact template but create a unique and engaging design for the content.
⟨
"slides": [
{{
"title": "",
"content": ""
⟩,
⟨
"title": "",
"content": "",
"notes": ""
⟩,
⟨
"title": "",
"content": "",
"notes": ""
⟩,
⟨
"title": "",
"content": "",
"notes": ""
⟩,
⟨
"title": "",
"content": "",
"notes": ""
⟩,
⟨
"title": "",
"content": "",
"notes": ""
⟩,
⟨
"title": "",
"content": "",
"notes": ""
⟩,
⟨
"title": "",
"content": "",
"notes": ""
⟩,
⟨
"title": "",
"content": "",
"notes": ""
⟩# The closing curly bracket for the last slide
] # Closing square bracket for slides array
}} # Closing curly bracket for JSON object
Additional Instructions:
✅ "content" should be short and structured for a slide Do not use bold formatting,
✅ "notes" should be detailed and explanatory, providing depth beyond the slide.
