IPA IP 2023 Question 74

A model that imitates the brain’s nervous system on a computer — a neural network is the model behind deep learning.

A neural network is built from artificial neurons (nodes) arranged in layers: an input layer, one or more hidden layers and an output layer. Each node takes the weighted sum of its inputs, passes it through an activation function and sends the result on to the next layer, much as a nerve cell fires to the cells it is connected to. Training adjusts the weights from data until the outputs come out right. Deep learning is a neural network with many hidden layers, which lets it find features such as the shapes in an image by itself.

Answer (d)

Why not others:

• •(a) — information devices everywhere in society that can connect to a network anytime and anywhere describes a ubiquitous network (ubiquitous computing)
• •(b) — the academic network linking universities and research institutions, run by Japan’s National Institute of Informatics, is SINET (Science Information NETwork)
• •(c) — local governments networked so that a resident’s name, date of birth, sex and address can be checked from other municipalities is Japan’s Basic Resident Registration Network (Juki Net)

Key rule: “Neural” points to the brain: a neural network models nerve cells and the strength of their connections, and it is the basis of deep learning; the other networks in such questions are communication networks.

Basic Theory · AI-written, may contain errors