ITPEC IP April 2025 Question 87

A computer learns the characteristics of data from a large volume of it with a model imitating human neural circuits — that is deep learning.

Deep learning is machine learning with a neural network of many layers. Each layer passes weighted values to the next, and training adjusts the weights so that the output matches the data. With many hidden layers, the network finds by itself the features that matter, for example the edges and shapes that identify a face, instead of a person designing them. It needs large volumes of data and computing power, and it drives image recognition, speech recognition and machine translation.

Answer (d)

Why not others:

• •(a) — integrating information and workflow among the departments that deal with customers (sales, marketing, after-sales service) is CRM (customer relationship management)
• •(b) — education and training delivered on computers and networks is e-learning
• •(c) — sharing the knowledge and know-how of individuals so the whole organization can use it is knowledge management

Key rule: Deep learning is machine learning with multilayer neural networks that extract the features of data by themselves; in classic machine learning, people choose the features.

Basic Theory · AI-written, may contain errors