Название: What AI Can Do: Strengths and Limitations of Artificial Intelligence
Автор: Manuel Cebral-Loureda, Elvira G. Rincon-Flores, Gildardo Sanchez-Ante
Издательство: CRC Press
Год: 2024
Страниц: 459
Язык: английский
Формат: pdf (true)
Размер: 17.6 MB

Most recent Artificial Intelligence (AI) advances have revolved around Machine Learning (ML) methods. Such methods can learn from data and create models that represent a problem (e.g., regression, classification). Traditional ML techniques are useful when working with structured data (i.e., tabular or grid data). During the process of ML, one essential step is to extract features from data, which are then used to learn and improve the model. This is generally done by manually designing the data to look for relationships across independent variables. However, when dealing with unstructured data (e.g., images, text, and sound), ML techniques struggle with extracting effective features from the data. This is due to the complexity of the data, which does not permit to extract sufficiently meaningful and powerful representations to feed and train the underlying ML models. Recent strides in AI have happened thanks to a series of factors that have enabled Deep Learning (DL) models come to the fore of ML research. In particular, DL models have shown an astounding capability for extracting highly efficient features and to learn very complex functions of the features for a great variety of talks, in an endto-end and automatic manner.