Machine Learning Specialization · Unsupervised Learning, Recommenders, Reinforcement Learning

Deep Q-Learning Lunar Lander

A Deep Q-Learning agent trained to land a LunarLander safely using rewards, a Q-network, a target network, and experience replay.

Reinforcement learningDeep Q-NetworkExperience replayTarget networkBellman update

GitHub

What this project is about

This is the reinforcement-learning milestone where value estimation becomes a neural network. The agent observes the lander state, chooses actions, receives rewards, and learns Q-values through replay and a target network.

Implementation focus

The strongest learning was the full RL loop: experience collection, Bellman targets, mini-batch training, target-network updates, and reward-based evaluation until the lander solves the environment.

Key results and artifacts

State shape: 8 features; actions: 4
Solved in 566 episodes
Last-100-episode average: 200.53

What I learned

This milestone turned reward feedback into policy learning and connected theoretical Q-values to a working control agent.

Milestone deep dive

What this project proves in my learning path.

A Deep Q-Learning agent trained to land a LunarLander safely using rewards, a Q-network, a target network, and experience replay. The milestone is included because it shows a complete learning unit: a problem, a data representation, an implementation path, measurable outputs, and a lesson that later connects to larger AI systems.

Input layerstate, action, reward, and next-state transitions

Model layervalue estimation or policy learning loop

Evaluation layertests, metrics, plots, or qualitative output

Artifact layerlearned action preferences and reward curves

Problem framing

What I was trying to solve

This page is not just a lab description. It explains how I transformed the assignment idea into an engineering story: what the model receives, what it optimizes, what output it should produce, and how the result can be inspected.

Implementation

What I coded or assembled

The implementation focus was the mechanism behind the result: functions, model blocks, training logic, preprocessing, evaluation, and the small technical details that make the notebook work rather than only look correct.

Concepts

Core concepts

Reinforcement learningDeep Q-NetworkExperience replayTarget networkBellman update

These tags show the technical ideas this milestone added to my toolkit.

Learning result

What changed after it

This milestone turned reward feedback into policy learning and connected theoretical Q-values to a working control agent.

Important outputs and evidence

State shape: 8 features; actions: 4
Solved in 566 episodes
Last-100-episode average: 200.53

How I would explain it in an interview

I would describe this milestone by starting from the data representation, then explain the model or algorithm, then show what was measured. The important part is not only the final number; it is the ability to reason about why the method works, what assumptions it makes, and what limitations should be improved next.