All Roadmaps / Artificial Intelligence

Artificial Intelligence Interactive Graph Curriculum

Artificial Intelligence

A comprehensive, modern roadmap to Artificial Intelligence — from intelligent agents and heuristic search to deep neural networks, transformer models, and autonomous tool-calling agents.

6 Phases

18 Topics

~324h Total Time

Study Progress: 0 of 18 completed • 0h of ~324h 0%

Status:

Level:

🎯

Recommended Next Step Beginner ⏱️ 3 hours

Intelligent Agents & Problem Formulation

Graph Legend:

Recommended

Optional

Beginner

Intermediate

Advanced

Done • In Progress • Use J/K or click cards to explore

Curriculum Start

01 Phase 01 0/3 completed

Phase 1: Foundations of Artificial Intelligence

Core concepts of intelligent rational agents, problem spaces, and search algorithms.

beginner Recommended • 1 week

Intelligent Agents & Environments

Understanding the PEAS framework (Performance measure, Environment, Actuators, Sensors), agent architectures, and environment properties (fully vs partially observable, deterministic vs stochastic).

intermediate Recommended • 10 days

State Space Search & Uninformed Search

Formulating problems as state spaces and finding paths to goals using Breadth-First Search (BFS), Depth-First Search (DFS), and Uniform Cost Search (UCS).

intermediate Recommended • 2 weeks

Heuristic & Adversarial Search (A* and Minimax)

Informed search algorithms using admissible heuristics (A* Search) and decision-making in competitive two-player zero-sum games (Minimax with Alpha-Beta pruning).

02 Phase 02 0/3 completed

Phase 2: Knowledge Representation & Logical Reasoning

Formal symbolic logic, ontological knowledge graphs, and handling uncertainty.

intermediate Recommended • 2 weeks

Propositional & First-Order Logic

Formal logic syntax and semantics, truth tables, inference rules (Modus Ponens, Resolution), unification, and forward/backward chaining.

intermediate Optional • 1 week

Knowledge Graphs & Semantic Ontologies

Entity-relation modeling, RDF triples, SPARQL querying, and modern vector-grounded enterprise knowledge graphs.

advanced Recommended • 2 weeks

Probabilistic Reasoning & Bayesian Networks

Modeling uncertainty with probability distributions, conditional independence, Bayes rule, and inference in Directed Acyclic Graphical Models (Bayesian Networks).

03 Phase 03 0/3 completed

Phase 3: Decision Making Under Uncertainty & Reinforcement Learning

Dynamic programming, Markov processes, and learning from trial-and-error reward signals.

intermediate Recommended • 2 weeks

Markov Decision Processes (MDP)

State spaces, action spaces, transition probabilities, discount factors, and solving for optimal policies via Bellman Equations and Value/Policy Iteration.

advanced Recommended • 2 weeks

Model-Free RL (Q-Learning & SARSA)

Temporal Difference learning without explicit transition dynamics. Exploration vs exploitation (epsilon-greedy), Q-value estimation, and off-policy vs on-policy learning.

advanced Optional • 3 weeks

Deep Reinforcement Learning (DQN & Policy Gradients)

Combining deep neural network function approximators with RL: Deep Q-Networks (experience replay, target networks), REINFORCE, and Actor-Critic (PPO).

04 Phase 04 0/3 completed

Phase 4: Machine Learning Core for AI Systems

Statistical learning theory, optimization algorithms, and predictive model pipelines.

beginner Recommended • 3 weeks

Supervised & Unsupervised Learning Foundations

Parametric vs non-parametric models, linear/logistic regression, support vector machines, decision trees, k-means clustering, and PCA dimensionality reduction.

intermediate Recommended • 10 days

Loss Formulation, Convexity & Gradient Descent

Mathematical foundations of optimization: gradient vectors, Hessian matrices, convex objectives, stochastic gradient descent (SGD), Adam, and learning rate schedules.

intermediate Recommended • 1 week

Feature Representation & Model Evaluation

Feature encoding, normalization, stratified k-fold validation, precision/recall trade-offs, ROC-AUC, and leakage prevention.

05 Phase 05 0/3 completed

Phase 5: Deep Learning & Frontier AI Architectures

Modern neural network topologies, backpropagation calculus, and transformer architectures.

intermediate Recommended • 2 weeks

Neural Networks & Backpropagation Calculus

Multi-layer perceptrons (MLP), matrix calculus, computational graphs, non-linear activation functions (ReLU, GELU), and backprop chain rule implementation.

advanced Recommended • 3 weeks

The Transformer Architecture & Self-Attention

Scaled dot-product attention, multi-head attention, positional encodings, layer normalization, residual connections, and encoder-decoder topologies.

advanced Recommended • 3 weeks

Large Language Models & Pre-training

Autoregressive next-token prediction, tokenization (BPE, WordPiece), scaling laws, distributed training (FSDP, Megatron), and instruction fine-tuning (SFT, RLHF, DPO).

06 Phase 06 0/3 completed

Phase 6: Autonomous Agents, RAG & AI Safety

Building production AI systems with external tools, knowledge grounding, and alignment.

advanced Recommended • 2 weeks

Autonomous Agent Frameworks & Tool Calling

ReAct prompting pattern (Reasoning + Acting), function calling, tool execution loops, planning, short-term and long-term memory architectures.

intermediate Recommended • 2 weeks

Retrieval-Augmented Generation (RAG)

Vector embeddings, chunking strategies, cosine similarity search, approximate nearest neighbors (HNSW), re-ranking, and grounded generation pipelines.

beginner Recommended • 1 week

AI Safety, Alignment & Robustness

Mechanistic interpretability, prompt injection defense, red teaming, constitutional AI, evaluation benchmarks, and ethical societal alignment.

Curriculum Breakdown

Phase 1: Foundations of Artificial Intelligence

0/3 completed

Core concepts of intelligent rational agents, problem spaces, and search algorithms.

Intelligent Agents & Environments

beginner ⏱️ 1 week

Understanding the PEAS framework (Performance measure, Environment, Actuators, Sensors), agent architectures, and environment properties (fully vs partially observable, deterministic vs stochastic).

Key Outcomes:

✓ Formulating agent-environment interaction cycles

✓ Differentiating simple reflex, goal-based, and utility-based agents

✓ Analyzing state observability and environmental complexity

Links: Russell & Norvig - AIMA Official Companion MIT 6.034: Artificial Intelligence Lecture 1

State Space Search & Uninformed Search

intermediate ⏱️ 10 days

Formulating problems as state spaces and finding paths to goals using Breadth-First Search (BFS), Depth-First Search (DFS), and Uniform Cost Search (UCS).

Key Outcomes:

✓ State space graph formulation

✓ Completeness, time complexity, and space complexity analysis

✓ Trade-offs between BFS, DFS, and Iterative Deepening

Links: Search Algorithms in AI (Stanford CS221) Graph Traversal & Search Foundations

Heuristic & Adversarial Search (A* and Minimax)

intermediate ⏱️ 2 weeks

Informed search algorithms using admissible heuristics (A* Search) and decision-making in competitive two-player zero-sum games (Minimax with Alpha-Beta pruning).

Key Outcomes:

✓ Heuristic admissibility and consistency conditions

✓ A* optimality guarantees: f(n) = g(n) + h(n)

✓ Game trees and Alpha-Beta pruning efficiency

Links: Introduction to A* Algorithm (Red Blob Games) Minimax and Alpha-Beta Pruning Explained

Phase 2: Knowledge Representation & Logical Reasoning

0/3 completed

Formal symbolic logic, ontological knowledge graphs, and handling uncertainty.

Propositional & First-Order Logic

intermediate ⏱️ 2 weeks

Formal logic syntax and semantics, truth tables, inference rules (Modus Ponens, Resolution), unification, and forward/backward chaining.

Key Outcomes:

✓ Representing real-world facts in First-Order Logic (FOL)

✓ Resolution theorem proving algorithm

✓ Horn clauses and rule-based expert systems

Links: Stanford Introduction to Logic First-Order Logic Notes (Berkeley CS188)

Knowledge Graphs & Semantic Ontologies

intermediate Optional ⏱️ 1 week

Entity-relation modeling, RDF triples, SPARQL querying, and modern vector-grounded enterprise knowledge graphs.

Key Outcomes:

✓ Semantic web standards (RDF, OWL, SPARQL)

✓ Extracting relations from unstructured text

✓ Hybrid graph-RAG architectures

Links: Knowledge Graphs (ACM Digital Library) Neo4j Graph Database Fundamentals

Probabilistic Reasoning & Bayesian Networks

advanced ⏱️ 2 weeks

Modeling uncertainty with probability distributions, conditional independence, Bayes rule, and inference in Directed Acyclic Graphical Models (Bayesian Networks).

Key Outcomes:

✓ Joint probability factorization with DAGs

✓ Variable elimination and exact inference

✓ Sampling methods (MCMC, Gibbs Sampling)

Links: Probabilistic Graphical Models (Coursera - Daphne Koller) Bayesian Reasoning and Machine Learning (David Barber)

Phase 3: Decision Making Under Uncertainty & Reinforcement Learning

0/3 completed

Dynamic programming, Markov processes, and learning from trial-and-error reward signals.

Markov Decision Processes (MDP)

intermediate ⏱️ 2 weeks

State spaces, action spaces, transition probabilities, discount factors, and solving for optimal policies via Bellman Equations and Value/Policy Iteration.

Key Outcomes:

✓ The Bellman Optimality Equation

✓ Value Iteration algorithm convergence

✓ Policy Iteration vs Value Iteration trade-offs

Links: Sutton & Barto: Reinforcement Learning Book David Silver RL Course (DeepMind/UCL)

Model-Free RL (Q-Learning & SARSA)

advanced ⏱️ 2 weeks

Temporal Difference learning without explicit transition dynamics. Exploration vs exploitation (epsilon-greedy), Q-value estimation, and off-policy vs on-policy learning.

Key Outcomes:

✓ TD(0) error calculation

✓ Q-Learning update rule: Q(s,a) <- Q(s,a) + alpha * [r + gamma * max Q(s',a') - Q(s,a)]

✓ Convergence guarantees under Robbins-Monro conditions

Links: OpenAI Spinning Up in Deep RL Q-Learning In-Depth Guide

Deep Reinforcement Learning (DQN & Policy Gradients)

advanced Optional ⏱️ 3 weeks

Combining deep neural network function approximators with RL: Deep Q-Networks (experience replay, target networks), REINFORCE, and Actor-Critic (PPO).

Key Outcomes:

✓ Overcoming deadly triad instability in Deep Q-Networks

✓ Policy gradient theorem and score functions

✓ Proximal Policy Optimization (PPO) clip objective

Links: Hugging Face Deep RL Course Playing Atari with Deep Reinforcement Learning (Mnih et al.)

Phase 4: Machine Learning Core for AI Systems

0/3 completed

Statistical learning theory, optimization algorithms, and predictive model pipelines.

Supervised & Unsupervised Learning Foundations

beginner ⏱️ 3 weeks

Parametric vs non-parametric models, linear/logistic regression, support vector machines, decision trees, k-means clustering, and PCA dimensionality reduction.

Key Outcomes:

✓ Formulating loss functions (MSE, Cross-Entropy)

✓ Bias-variance decomposition

✓ Regularization techniques (L1 Lasso, L2 Ridge)

Links: EncodeEdge Introduction to Machine Learning Scikit-Learn User Guide

Loss Formulation, Convexity & Gradient Descent

intermediate ⏱️ 10 days

Mathematical foundations of optimization: gradient vectors, Hessian matrices, convex objectives, stochastic gradient descent (SGD), Adam, and learning rate schedules.

Key Outcomes:

✓ Computing analytical and numerical gradients

✓ Momentum, RMSprop, and Adam optimizer equations

✓ Saddle points, local minima, and loss surface geometry

Links: 3Blue1Brown: Gradient Descent & Neural Networks An Overview of Gradient Descent Optimization Algorithms

Feature Representation & Model Evaluation

intermediate ⏱️ 1 week

Feature encoding, normalization, stratified k-fold validation, precision/recall trade-offs, ROC-AUC, and leakage prevention.

Key Outcomes:

✓ Data preprocessing pipelines without data leakage

✓ Choosing error metrics for imbalanced datasets

✓ Cross-validation and confidence intervals

Links: EncodeEdge Regression Error Metrics Guide Google Machine Learning Crash Course

Phase 5: Deep Learning & Frontier AI Architectures

0/3 completed

Modern neural network topologies, backpropagation calculus, and transformer architectures.

Neural Networks & Backpropagation Calculus

intermediate ⏱️ 2 weeks

Multi-layer perceptrons (MLP), matrix calculus, computational graphs, non-linear activation functions (ReLU, GELU), and backprop chain rule implementation.

Key Outcomes:

✓ Deriving backpropagation for a two-layer neural net

✓ Vanishing and exploding gradient mitigation

✓ PyTorch autograd mechanics

Links: Andrej Karpathy: Building micrograd from scratch PyTorch Deep Learning Zero to Mastery

The Transformer Architecture & Self-Attention

advanced ⏱️ 3 weeks

Scaled dot-product attention, multi-head attention, positional encodings, layer normalization, residual connections, and encoder-decoder topologies.

Key Outcomes:

✓ Attention formula: Attention(Q, K, V) = softmax(QK^T / sqrt(d_k))V

✓ Computational complexity of self-attention: O(N^2)

✓ Causal masking in decoder-only models (GPT)

Links: Attention Is All You Need (Vaswani et al.) The Illustrated Transformer by Jay Alammar

Large Language Models & Pre-training

advanced ⏱️ 3 weeks

Autoregressive next-token prediction, tokenization (BPE, WordPiece), scaling laws, distributed training (FSDP, Megatron), and instruction fine-tuning (SFT, RLHF, DPO).

Key Outcomes:

✓ Next-token prediction cross-entropy loss

✓ Byte Pair Encoding (BPE) tokenization implementation

✓ Aligning pre-trained base models with human instructions

Links: Let's build GPT: from scratch, in code (Andrej Karpathy) Hugging Face Transformers Documentation

Phase 6: Autonomous Agents, RAG & AI Safety

0/3 completed

Building production AI systems with external tools, knowledge grounding, and alignment.

Autonomous Agent Frameworks & Tool Calling

advanced ⏱️ 2 weeks

ReAct prompting pattern (Reasoning + Acting), function calling, tool execution loops, planning, short-term and long-term memory architectures.

Key Outcomes:

✓ ReAct loop implementation (Thought -> Action -> Observation -> Answer)

✓ Structured output schema enforcement with JSON

✓ State management in multi-step autonomous workflows

Links: ReAct: Synergizing Reasoning and Acting in Language Models LangChain / LangGraph Agent Documentation

Retrieval-Augmented Generation (RAG)

intermediate ⏱️ 2 weeks

Vector embeddings, chunking strategies, cosine similarity search, approximate nearest neighbors (HNSW), re-ranking, and grounded generation pipelines.

Key Outcomes:

✓ Document ingestion, chunking, and embedding generation

✓ Vector database indexing (Qdrant, Milvus, Chroma)

✓ Hybrid dense + sparse BM25 retrieval and Cross-Encoder re-ranking

Links: LlamaIndex Documentation Pinecone Learning Center: What is Vector Search?

AI Safety, Alignment & Robustness

beginner ⏱️ 1 week

Mechanistic interpretability, prompt injection defense, red teaming, constitutional AI, evaluation benchmarks, and ethical societal alignment.

Key Outcomes:

✓ Vulnerabilities: Prompt injection, jailbreaks, and hallucinations

✓ Guardrails, output moderation, and deterministic evaluators

✓ Ethical fairness, privacy, and alignment frameworks

Links: Anthropic Research on AI Safety & Alignment OpenAI Safety & Alignment Guidelines

Beginner 45 min

Topic Title

Status:

What You Need to Know

Topic explanation description goes here.

Code Blueprint

Progress exported & copied to clipboard!

Artificial Intelligence

Intelligent Agents & Problem Formulation

Phase 1: Foundations of Artificial Intelligence

Intelligent Agents & Environments

State Space Search & Uninformed Search

Heuristic & Adversarial Search (A* and Minimax)

Phase 2: Knowledge Representation & Logical Reasoning

Propositional & First-Order Logic

Knowledge Graphs & Semantic Ontologies

Probabilistic Reasoning & Bayesian Networks

Phase 3: Decision Making Under Uncertainty & Reinforcement Learning

Markov Decision Processes (MDP)

Model-Free RL (Q-Learning & SARSA)

Deep Reinforcement Learning (DQN & Policy Gradients)

Phase 4: Machine Learning Core for AI Systems

Supervised & Unsupervised Learning Foundations

Loss Formulation, Convexity & Gradient Descent

Feature Representation & Model Evaluation

Phase 5: Deep Learning & Frontier AI Architectures

Neural Networks & Backpropagation Calculus

The Transformer Architecture & Self-Attention

Large Language Models & Pre-training

Phase 6: Autonomous Agents, RAG & AI Safety

Autonomous Agent Frameworks & Tool Calling

Retrieval-Augmented Generation (RAG)

AI Safety, Alignment & Robustness

Curriculum Breakdown

Phase 1: Foundations of Artificial Intelligence

Intelligent Agents & Environments

State Space Search & Uninformed Search

Heuristic & Adversarial Search (A* and Minimax)

Phase 2: Knowledge Representation & Logical Reasoning

Propositional & First-Order Logic

Knowledge Graphs & Semantic Ontologies

Probabilistic Reasoning & Bayesian Networks

Phase 3: Decision Making Under Uncertainty & Reinforcement Learning

Markov Decision Processes (MDP)

Model-Free RL (Q-Learning & SARSA)

Deep Reinforcement Learning (DQN & Policy Gradients)

Phase 4: Machine Learning Core for AI Systems

Supervised & Unsupervised Learning Foundations

Loss Formulation, Convexity & Gradient Descent

Feature Representation & Model Evaluation

Phase 5: Deep Learning & Frontier AI Architectures

Neural Networks & Backpropagation Calculus

The Transformer Architecture & Self-Attention

Large Language Models & Pre-training

Phase 6: Autonomous Agents, RAG & AI Safety

Autonomous Agent Frameworks & Tool Calling

Retrieval-Augmented Generation (RAG)

AI Safety, Alignment & Robustness

No matching topics found

Topic Title

What You Need to Know

Code Blueprint

Key Milestones & Checkpoints

Recommended Prerequisites

Personal Study Notes

Video Tutorial

Official Guides & Documentation

Further Reading & Standards

Related EncodeEdge Tutorials

Keyboard Shortcuts

Reset All Progress?