Public · self-paced · free
AI for Beginners
Build a clear, practical foundation in artificial intelligence—one lesson at a time.
A practical, self-paced path from the foundations of artificial intelligence to neural networks, language, vision, responsible AI, and modern multimodal systems.
- Core lessons
- 24
- Estimated time
- 34 hours
- Labs
- 12
- Languages
- Najdi Arabic · English
Your progress
0%0 of 24 core lessons complete
Without sign-in, progress stays in this browser.
Start the courseCurriculum
Your learning path
Follow the modules in order, or open any lesson. All 24 core lessons are required for course completion; the final lesson is a bonus.
Module 1
Foundations
Module 2
Neural Networks
- Not started03Introduction to Neural Networks: PerceptronCalculate a perceptron output from its inputs, weights, and bias.75 min2 labs
- Not started04Introduction to Neural Networks. Multi-Layered PerceptronSelect an appropriate loss function for a regression or classification task.90 min2 labs
- Not started05Neural Network FrameworksCompare the model-building and training workflows in PyTorch, TensorFlow, and Keras.90 min4 labs
Module 3
Computer Vision
- Not started06Introduction to Computer VisionLoad images as numeric arrays and apply resizing, color, blur, and threshold operations with OpenCV.75 min2 labs
- Not started07Convolutional Neural NetworksExplain how convolutional filters, feature maps, and pooling preserve useful spatial patterns.95 min3 labs
- Not started08Pre-trained Networks and Transfer LearningReuse a pre-trained vision network as a frozen feature extractor or a fine-tuned model.95 min5 labs
- Not started09AutoencodersTrace an image through an encoder, latent representation, and decoder.80 min2 labs
- Not started10Generative Adversarial NetworksDistinguish the generator and discriminator objectives in adversarial training.90 min4 labs
- Not started11Object DetectionDistinguish image classification, localization, and multi-object detection outputs.85 min2 labs
- Not started12SegmentationFormulate semantic segmentation as a class prediction for every image pixel.85 min3 labs
Module 4
Natural Language Processing
- Not started13Representing Text as TensorsTokenize text at character and word level and map the tokens into tensors.75 min2 labs
- Not started14EmbeddingsExplain how dense embedding vectors encode useful semantic relationships between words.80 min2 labs
- Not started15Language ModelingDescribe a language model as a probability distribution over the next token in context.85 min3 labs
- Not started16Recurrent Neural NetworksTrace inputs and hidden state through a recurrent neural-network cell.90 min2 labs
- Not started17Generative networksMap one-to-one, one-to-many, and sequence-to-sequence recurrent architectures to suitable tasks.90 min3 labs
- Not started18Attention Mechanisms and TransformersCalculate how attention weights combine token representations and context.100 min2 labs
- Not started19Named Entity RecognitionFormulate named-entity recognition as token classification with labels for multi-token entities.75 min2 labs
- Not started20Pre-Trained Large Language ModelsExplain how pre-trained transformer language models support generation and zero-shot task adaptation.80 min1 labs
Module 5
Other AI Techniques
- Not started21Genetic AlgorithmsEncode candidate solutions as genes and define a fitness function for a search problem.70 min2 labs
- Not started22Deep Reinforcement LearningModel a reinforcement-learning problem with states, actions, rewards, and an environment.95 min5 labs
- Not started23Multi-Agent SystemsDescribe an agent through its environment, observations, actions, and goals.70 min
Module 6
Responsible AI
Module 7
Bonus
Before you begin
Start here
This course is a native learning experience inside the site: lessons, assessments, labs, search, and progress work together without cloning a separate repository.
The lesson rhythm
- Review the objectives, prerequisites, and estimated time.
- Try the diagnostic pre-check; “I’m not sure” is always available and there is no pass gate.
- Read the lesson and use the notebook or lab that matches your framework track.
- Submit the post-check, then review the rationale and linked lesson section for every answer.
- Explicitly mark the lesson complete.
Choose a framework
Select PyTorch or TensorFlow and switch later without losing lesson completion. Framework-neutral notebooks remain available in both tracks.
Progress and privacy
Progress works locally without sign-in. If you sign in, completed work and best scores merge safely with your account for cross-device resume. The platform stores completion states and scores, never individual answers or assignment and notebook content.
Course completion
The core course contains 24 lessons; lesson 25 is optional bonus material. Labs, assignments, and scores do not gate completion, and the first release does not issue a certificate.
Run notebooks safely
Every lab includes a sanitized, read-only notebook preview inside the site. Execution results, errors, counters, widgets, and active output are removed before publication.
Recommended workflow
- Review the explanation and code in the native preview.
- Download the clean
.ipynbfile from the lab. - Run it in an isolated Python environment, or use Colab or Codespaces when an approved immutable edition exposes those actions.
- Install only the packages imported by that notebook and pin them in your own project before relying on the result.
Local execution
Use a fresh virtual environment for each experiment. PyTorch, TensorFlow, computer-vision, and NLP labs have different dependency needs, so inspect the first import cells before installing packages. Never place API keys or personal data in a notebook.
External execution
Colab and Codespaces actions appear only when the embedded source has an immutable Git release identity. External services have their own privacy terms, quotas, and costs; review them before uploading data.