D802: Deep Learning

Deep Learning goes underneath the black box — the actual mathematics and mechanics of gradient descent, backpropagation, and the neural network architectures that power modern AI.

What D802 covers

The course delves into fundamental principles, underlying mathematics, and implementation details of deep learning, covering gradient descent, backpropagation, and computation graphs.

Students explore modules constituting deep learning models — linear, convolution, and pooling layers, activation functions — and common architectures including CNNs and RNNs, gaining skills to design, implement, and optimize advanced deep learning systems.

The D802 performance assessment

Expect a performance assessment requiring you to design and implement a deep learning model (CNN or RNN) for a given problem, explaining the architecture choices and optimization approach.

Key topics in D802

• Gradient descent and backpropagation
• Convolutional and pooling layers
• CNN and RNN architectures
• Deep learning system optimization

Writing tips for D802

Follow the task instructions and rubric line by line

WGU performance assessments for D802 are graded against a fixed rubric — every rubric line has to be visibly addressed, usually with a labeled heading that mirrors the rubric language. Skipping a rubric point because it seems minor is the single most common reason a competent submission comes back "Not Yet Competent" for revision.

Show your work: code, reasoning, and test results, not just a final answer

WGU evaluators are trained to distinguish genuine technical work from a paraphrased summary. Include your actual code, algorithmic reasoning, and test/benchmark results, not just a description of what you built — a rubric checking technical competency wants to see the artifact and the thinking behind it.

Because WGU is self-paced, don't let "no deadline pressure" become no submission

There's no weekly due date forcing progress, which means procrastination costs more at WGU than at a traditional term-based school — a stalled task can quietly eat weeks of a term. Treat your own target date for each D802 assessment as a real deadline.

Why students seek help with D802

Students sometimes use a neural network architecture without justifying why that specific architecture (CNN vs. RNN) fits the problem — the course wants that architectural reasoning explicit.

How GradeEssays helps with D802

Share your deep learning problem and rubric, and your writer will build the model implementation with clear justification for the architecture and optimization choices made.

Prerequisites and program context

D802 requires Machine Learning for Computer Scientists (D801) as a prerequisite.

Related courses