ITPEC IP April 2026 Question 13

Neural network — the model of connected artificial neurons on which deep learning is built.

A neural network connects simple units (artificial neurons) in layers. Each unit multiplies its inputs by weights, adds them up and passes the result to the next layer. Training adjusts the weights so that the outputs match the training data. Deep learning uses networks with many hidden layers, which learn the features of large volumes of data, such as images or speech, by themselves.

Answer (b)

Why not others:

• •(a) Social network — the relationships among people, as on an SNS; it is not a learning model
• •(c) Feasibility study — a planning-stage study of whether a project is technically and economically possible
• •(d) Fault tolerant — a system design that keeps working when part of it fails, through redundancy

Key rule: Deep learning is machine learning with multilayer neural networks; "deep" refers to the many hidden layers, which extract features without a person defining them.

Basic Theory · AI-written, may contain errors