Huawei H13-311: HCIA-AI V4.0

Last checked November 10, 2026 against the official Huawei exam page.

1. Question 1Basics of Deep Learning and Foundation Models

   A student builds a single perceptron to learn the XOR function. Despite many training epochs, the perceptron never classifies all four input patterns correctly. What best explains this result?

   1. AThe learning rate was set too high, which stops any perceptron from learning binary functions of any kind
   2. BPerceptrons accept only continuous numeric inputs, so binary XOR inputs cannot be processed by the model
   3. CXOR is not linearly separable, so one perceptron cannot draw the single straight boundary that the task needs
   4. DThe perceptron uses a softmax output layer, which is unable to represent exactly two output classes at once
   Show answer and explanationHide answer

   Correct answer: C

   XOR outputs 1 only when the two inputs differ, so no single straight line separates the classes, and a single perceptron is limited to linearly separable problems. A learning rate problem would not block all binary functions, perceptrons accept numeric inputs including binary ones, and perceptrons typically use a step activation rather than softmax.

2. Question 2Machine Learning Overview

   A telecom company has historical records in which every customer is already tagged as either cancelled or retained. It wants a model that estimates which current customers are likely to cancel. Which learning paradigm best fits this requirement?

   1. AUnsupervised learning, using clustering to discover hidden groups among customers without any labels
   2. BSupervised learning, using the labeled historical records to train a classifier that predicts the outcome
   3. CReinforcement learning, using a reward signal that the system receives after each customer decision
   4. DDimensionality reduction, using principal components to compress the customer data into fewer features
   Show answer and explanationHide answer

   Correct answer: B

   Labeled historical outcomes paired with a target to predict define supervised learning, here a classification task. Clustering ignores the existing labels, reinforcement learning needs an agent learning from rewards over time, and dimensionality reduction does not predict an outcome.

3. Question 3AI Development Framework

   Which characteristic best describes PyTorch's default execution model?

   1. AA static computation graph that must be fully compiled before any training data is processed
   2. BCPU-only execution, with no support for GPU or other hardware accelerators at all
   3. CA dynamic computation graph that is built as operations run, which makes debugging easier
   4. DA mandatory conversion of every model into a separate exchange format before it can be trained
   Show answer and explanationHide answer

   Correct answer: C

   PyTorch builds its graph dynamically while the code executes, so developers can inspect intermediate values and debug with ordinary Python tools. Static graphs are a different execution style, PyTorch supports GPU acceleration, and models do not need to be converted before training.

4. Question 4AI Business Process Overview

   A bank wants to use AI to reduce the number of loans that default. No data has been gathered and no model has been chosen. Which step should come first in the AI business process?

   1. ADeploy a finished model to every branch so that staff can begin using it right away for decisions
   2. BDefine the business problem, the decision it supports, and how success will be measured
   3. CSelect the most complex deep learning architecture available so that the model captures every pattern
   4. DPurchase a large cluster of GPU servers before any requirements are written down or agreed upon
   Show answer and explanationHide answer

   Correct answer: B

   An AI business process starts by framing the problem, the decision it supports and the success metrics, which then guide data and model choices. Deployment, architecture selection and hardware purchases all come before the requirements are known, which risks wasted effort.

5. Question 5AI Overview

   Which statement correctly describes the relationship between artificial intelligence, machine learning and deep learning?

   1. ADeep learning is the broadest field, containing machine learning, which in turn contains artificial intelligence as a subset
   2. BArtificial intelligence is the broadest field, machine learning is a subset of it, and deep learning is a subset of machine learning
   3. CMachine learning and deep learning are separate fields that apply only to robotics and physical control systems in industry
   4. DArtificial intelligence is a narrow subset of machine learning that is used mainly for image recognition tasks in research labs
   Show answer and explanationHide answer

   Correct answer: B

   Artificial intelligence is the umbrella field of building systems that perform intelligent tasks, machine learning is a subset that learns from data, and deep learning is a subset of machine learning built on multi-layer neural networks. The other options reverse the hierarchy or wrongly narrow the scope of AI and machine learning to robotics or image recognition.