← ALL COURSES
CNSLT / ACADEMY / TRACK 1
FREE TEXT COURSE · OPEN ACCESS

AI Foundations: From First Principles to Neural Networks

A beginner pathway that builds technical understanding through a running spam-filter example.

BEGINNER10 PUBLISHED LESSONSSELF-PACEDNO LOGIN REQUIRED
START LEARNING Start with lesson 01 · Free text access
CHOOSE YOUR LEARNING FORMAT

Learn in the format that suits you.

AVAILABLE NOW

Free Text

Read the published lessons, work through the examples and use the knowledge checks at your own pace.

  • 10 published lessons
  • Worked examples and self-checks
  • Progress saved on this device
OPEN ACCESS · ₹0
VIDEO PRO · PREMIUM FORMAT

Guided video learning

A guided video-learning format for learners who prefer narrated explanations, visual walkthroughs and a more structured learning experience. Contact CNSLT for current access details.

  • Narrated concept explanations
  • Visual demonstrations and worked walkthroughs
  • Guided progression through the curriculum
ENQUIRE ABOUT VIDEO PRO PREMIUM LEARNING FORMAT
YOUR LEARNING PROGRESSLoading your progress…
0% complete · Saved on this device
PUBLISHED LEARNING PATH

Start with the lessons. Build in sequence.

These are the ten lessons currently published. The remaining Track 1 areas are retained in the roadmap below and will be released in verified batches.

MODULE 01

Foundations

Understand AI, machine learning and deep learning before examining a neural network.

MODULE 02

Inside a neuron

Build the arithmetic model step by step: inputs, weights, bias, weighted sum and activation.

PRACTICE / WORKED EXAMPLES

Follow the neuron calculation.

Example 1

Inputs: (1, 1, 1)

z = 1.5

sigmoid(z) ≈ 0.818

Example 2

Inputs: (1, 0, 1)

z = 0.6

sigmoid(z) ≈ 0.646

Example 3

Inputs: (0, 0, 0)

z = -0.5

sigmoid(z) ≈ 0.378

MASTERY EXERCISE

Inputs (1, 1, 0); weights (0.6, 0.7, 0.4); bias -0.3. Calculate the weighted sum, add the bias, apply sigmoid, then apply the illustrative 0.70 threshold.

Reveal expected calculation

Weighted input sum = 1.3; z = 1.0; sigmoid(1.0) ≈ 0.731; classify as spam because 0.731 > 0.70.

Outstanding — the learner has not submitted answers. This is not a passed assessment.

THE FULL TRACK 1 CURRICULUM

21 areas. Released in verified batches.

The first four processed concepts align with the first four areas in the 21-area curriculum. Concepts 5–10 in this release (inputs and features, weights, bias, weighted sum, activation functions, and model output/thresholds) are subtopics within neural-network foundations; they are not additional top-level curriculum areas.

  1. 01Artificial IntelligenceCONTENT PROCESSED
  2. 02Machine LearningCONTENT PROCESSED
  3. 03Deep LearningCONTENT PROCESSED
  4. 04Neural NetworksCONTENT PROCESSED
  5. 05Large Language ModelsPLANNED
  6. 06Tokens and TokenisationPLANNED
  7. 07EmbeddingsPLANNED
  8. 08TransformersPLANNED
  9. 09AttentionPLANNED
  10. 10Model TrainingPLANNED
  11. 11Model InferencePLANNED
  12. 12Model BehaviourPLANNED
  13. 13Model FamiliesPLANNED
  14. 14AI APIsPLANNED
  15. 15GPUs and Inference EconomicsPLANNED
  16. 16AI AgentsPLANNED
  17. 17Tools and Tool CallingPLANNED
  18. 18Model Context Protocol (MCP)PLANNED
  19. 19MultimodalityPLANNED
  20. 20Model EvaluationPLANNED
  21. 21Model SelectionPLANNED