Modern Deep Learning Foundations
A structured, evidence-centered route through six deep-learning concepts. Work with real or transparent data, inspect every result, and finish each laboratory with a reusable mastery artifact.
Modern Deep Learning Foundations
Learn by predicting, running mathematically grounded experiments, explaining evidence, and exporting an inspectable mastery artifact.
- Laboratories
- 6
- Guided time
- 4h 20m
- Mastery artifacts
- 6
A complete learning cycle, not a collection of demos
Read the relevant book passage, check the prerequisite, and make a testable prediction.
Change one variable at a time, capture reproducible runs, and inspect the displayed calculation.
Use the values on screen to explain the result, then apply the idea to a new technical decision.
Export a mastery artifact and schedule retrieval practice for durable understanding.
Your six-laboratory learning path
Understanding Evaluation with AUC-ROC Curve
Inspect the confusion matrix, move the classification threshold, and connect precision, recall, F1, ROC, and AUC on one consistent dataset.
- Level
- Foundation
- Guided time
- 35 min
- Confusion matrix
- Threshold analysis
- ROC and AUC
Overfitting and Regularization
Compare training and validation behaviour while model capacity, regularization strength, and data conditions change.
- Level
- Foundation
- Guided time
- 45 min
- Bias and variance
- Regularization
- Generalization gap
How Does a CNN Work?
Explore a complete CNN pipeline with real benchmark samples, editable kernels, padding, stride, pooling, and layer-level activations.
- Level
- Intermediate
- Guided time
- 50 min
- MNIST and CIFAR-10
- Activations and pooling
- Tensor shapes
Self-Attention and the Transformer Principle
Trace token embeddings through Query, Key, Value projections, scaled dot-product attention, multi-head mixing, and residual updates.
- Level
- Intermediate
- Guided time
- 45 min
- Q, K, and V
- Multi-head attention
- Residual stream
Explainability: Understanding Model Decisions
Compare local feature attribution methods and inspect how evidence supports or opposes a model prediction.
- Level
- Intermediate
- Guided time
- 40 min
- Feature attribution
- Local explanations
- Explanation stability
Transfer Learning and Fine-Tuning
Freeze and unfreeze a pretrained network block by block, then observe trainable parameters, adaptation, expected accuracy, and forgetting risk.
- Level
- Advanced
- Guided time
- 45 min
- Frozen layers
- Partial fine-tuning
- Catastrophic forgetting
Six inspectable artifacts define completion
A laboratory is complete only after you capture a reproducible run, answer the prediction and explanation prompts, and demonstrate transfer to a new context.
- 6 prerequisite checks
- 6 reproducible experiment records
- 6 evidence-based explanations
- 6 mastery artifacts with retention plans