The Math Behind AI for Filmmaking & Content Creation - Batch 2

The Math Behind AI for Filmmaking & Content Creation – Batch 2

₹500.00

Last date of registration : 7th January, 2026

90 in stock

SKU: IIT Madras | Date : 10th January, 2026 Categories: , ,

Description of the Workshop

To help filmmakers, artists, and creators understand the math and architecture behind the AI tools they use daily (like Midjourney, Runway, ChatGPT, etc.), including how these models learn, generate visuals, or create text.  Focus: Understand the core math & model structures behind visual & creative AI (Diffusion, GANs, Transformers) in simple, intuitive language

Profile of the Instructor

Mr. Shashwat Mookherjee is the Co-founder and CEO of IndieRise Research Labs Pvt. Ltd., a deep-tech content studio incubated at IIT Madras Research Park, aimed at revolutionizing filmmaking through AI-driven tools. A filmmaker, technologist, and creative producer, he directed the Marche Du film Cannes featured independent film Dispersion. He is a graduate of Data Science and Applications from IIT Madras.

MODULE 1

MODULE 1 — AI in Creativity: What’s Under the Hood? (10:00AM – 10:45 AM)  Topics Covered: 
  • What does an AI model actually do? 
  • What are weights, tokens, vectors, embeddings?
  • Difference between training and inference 
  • Why AI “learns” using optimization, not memorization 
Visual Examples: 
  • Use ChatGPT prompt → embedding visual → output path 
  • How a prompt like "epic wide shot of warrior" becomes pixels 

MODULE 2

MODULE 2 — Math of Diffusion Models (10:45AM – 11:45 AM) Topics Covered: 
  • What are diffusion models? Why are they used for visuals? 
  • Gaussian noise, timestep removal 
  • How Stable Diffusion and Midjourney actually generate images
  • Loss functions, denoising, latent spaces 
Visual Intuition Tools: 
  • Show frame-by-frame noise removal 
  • 3D latent space maps (e.g., seed variation) 

MODULE 3

MODULE 3 — Transformers for Text-to-Anything (11:45AM – 1:00 PM) Topics Covered: 
  • What is a transformer? Why is it so powerful? 
  • Attention mechanism (the math behind “focus”) 
  • GPT, ChatGPT, LLaMA: token → vector → output 
  • Prompt design as mathematical weight shaping
Visualizations: 
  • Show attention heatmaps 
  • Vector math for token influence (why “cinematic” matters more than “beautiful”) 

MODULE 4

MODULE 4 — Training & Fine-Tuning (3:00PM – 4:00 PM) Topics Covered: 
  • What happens during training? (Weights, epochs, backpropagation)
  • Finetuning types: Full, DreamBooth, LoRA 
  • Overfitting vs generalization 
  • Cost of training vs inference: why it’s slow and expensive 
Visual Breakdown: 
  • Show a tiny 2-layer network updating weights 
  • LoRA visual: how adding identity to model works 

MODULE 5

MODULE 5 — Bias, Errors & Ethics in AI (4:00PM – 5:00 PM) Topics Covered: 
  • Why AI “hallucinates” and what that means 
  • Dataset bias: Why Midjourney outputs westernized faces 
  • Copyright, dataset scraping, and moral gray zones
  • Fair use vs ethical use in filmmaking 
  • Open-source vs closed models 
Real-World Case Studies: 
  • Midjourney banning prompts 
  • Hollywood writer strikes over GPT 
  • Volatile's approach to ethical AI 

Eligibility & Intended Audience

Eligibility : The Workshop is Open to all

Fee for the Workshop

Registration fee : Rs. 590 (500 + 18% GST)

Pre - Requisite

Requirements 
  • No coding experience needed (but Python notebooks provided for visual learners)
  • Curiosity to understand AI beyond black-box use 
  • A notebook for doodling math concepts and mental models! 

Session Details

MODULES TIMINGS
MODULE 1 — AI in Creativity: What’s Under the Hood? 10:00AM – 10:45 AM
MODULE 2 — Math of Diffusion Models 10:45AM– 11:45 AM
MODULE 3 — Transformers for Text-to-Anything 11:45AM – 1:00 PM
MODULE 4 — Training & Fine-Tuning 3:00PM – 4:00 PM
MODULE 5 — Bias, Errors & Ethics in AI 4:00PM – 5:00 PM

Deliverables

  • Visual guide PDF: Diffusion, Transformer, LoRA, etc. 
  • Summary sheet of core math concepts (loss, gradients, latent) 
  • Demo Colab notebooks (optional for those who want to try math live)
  • Certificate of Completion 

Included Resources

  • Concept posters: Diffusion timeline, Transformer flow, LoRA layering
  • Dataset & Model Terminology reference chart 
  • Further reading links & videos for self-paced mastery 
  • Bonus: Explainer scripts for pitching AI tools to non-technical producers/stakeholders 

Learning Outcomes

By the end of the day, students will be able to:
  • Explain how diffusion and transformer models function — clearly, without code
  • Understand how AI models are trained, finetuned, and optimized
  • Identify and critique the ethical and creative limitations of AI 
  • Build confidence to work with AI as a creative partner — not just a black-box generator

Certification

Attending the workshop is mandatory for certification

Reviews

There are no reviews yet.

Be the first to review “The Math Behind AI for Filmmaking & Content Creation – Batch 2”