Chapter 3 of 6 · Huawei HCIA-AI H13-311
Basics of Deep Learning and Foundation Models
Perceptrons, neural network training, CNNs, RNNs and LSTMs, Transformers and foundation models.
- 3topics
- 17 minfast track
- 2diagrams and exhibits
- 30%of the exam
Topics in this chapter
Take them in order, or jump to the one you need.
- 1Perceptron and Fully-connected Neural NetworksLearn the perceptron and XOR, activation and loss functions, forward and backpropagation, SGD and Adam, vanishing gradients…5 min
- 2CNN and RNN ArchitecturesLearn convolution, pooling, output size and parameter counts in CNNs, then recurrent networks, vanishing gradients and the gates…Diagram6 min
- 3Transformer and Foundation Model InfrastructureLearn self-attention, multi-head attention, positional encoding, BERT and GPT, pre-training, fine tuning and LoRA, and compute…Diagram6 min
About this chapter
What this chapter covers
Basics of Deep Learning and Foundation Models is 30% of the Huawei HCIA-AI H13-311 exam, the largest chapter. It covers the perceptron, fully-connected networks and how they are trained, convolutional and recurrent networks, the Transformer, and the infrastructure behind large foundation models.
Questions are single-answer, multiple-answer without a stated count, true/false and typed fill-in-the-blank, often "Which of the following statements is true?". This exam has no exhibit figures and no drag and drop, so every question is plain text, and output-size and parameter-count calculations appear.
It has 6 outline topics, and three pages cover them.
How the topics fit together
A perceptron is one artificial neuron; stacking layers of them gives a fully-connected network trained by backpropagation. CNNs specialize this for images, RNNs for sequences, and the Transformer replaces recurrence with attention, which is what foundation models are built on.
PyTorch, covered in the AI Development Framework chapter, is how all of these are built in code.
Memorize the convolution output size formula, (W - K + 2P) / S + 1, and how to count parameters of a convolution and a linear layer. These small calculations are easy marks.
How to use these pages
Each topic page has the same shape, so you can skim for what you need.
- What it is: the idea in plain words.
- How it works: the mechanism, with tables and examples.
- What to remember for the exam: the facts to memorize.
- Common mistakes: the classic traps.
- Check yourself: three questions to test your understanding.
For each architecture you meet, say what kind of data it suits and what problem of the previous architecture it solves. The exam often asks why one design replaced another.
Ready for more practice?
The full Huawei HCIA-AI H13-311 question bank has 600 questions, with timed practice exams and an explanation for every answer.
This study guide is independent and is not affiliated with, endorsed by, or sponsored by Huawei. Vendor names, certification names and exam codes are trademarks of their respective owners and are used for identification only.