問95
頻出度★★☆
AI の関連技術であるディープラーニングに用いられる技術として, 最も適切なものはどれか。
令和6年度 ITパスポート試験 問95
テクノロジ系基礎理論 › 情報に関する理論
This question's explanation and your answers so far.
解説
問 95
Incorrectア
A social network refers to a social-type network — a mechanism (such as an SNS) that lets people connect and interact with one another over the Internet — and has nothing to do with AI learning methods. Its name resembles “neural network”, which is confusing, but it is not a deep learning technology.
- イ
正解 Deep learning is a machine learning method that builds a “neural network” — a mathematical model of how the nerve cells (neurons) in the human brain are connected — by stacking it into many layers: an input layer, intermediate (hidden) layers, and an output layer. By making the intermediate layers multilayered, it can automatically extract complex features of images, audio, natural language, and so on in a step-by-step manner. Therefore, the most appropriate technology used in deep learning is the neural network in choice イ.
Incorrectウ
A feasibility study refers to investigating and examining in advance the feasibility (in technical, cost-effectiveness, and other respects) of a system development effort or project before it is started. It is a project management term, not a deep learning technology.
Incorrectエ
Fault tolerant refers to a reliability design concept in which, even if a failure (fault) occurs in part of the system, the system as a whole maintains its functions and keeps operating. It is a term related to system availability, not a deep learning technology.
ここがポイント
- Deep learning is a representative machine learning method that stacks a neural network modeled on the brain’s neural circuits into multiple layers to automatically extract features.
- This question tests whether you can distinguish “neural network” from “social network”, as well as from terms in other fields that sound similar, such as “feasibility study” and “fault tolerant”.
用語