Dive into deeplearning Note
1. Introduction
Machine learning is the study of algorithms that can learn from experience.
We cannot formulate the process to obtain the output, but we know what the desired output should resemble.
Often, even when we do not know how to tell a computer explicitly how to map from inputs to outputs, we are nonetheless capable of performing the cognitive feat ourselves.
1.2 Key Components
This sort of problem, where we try to predict a designated unknown label based on known inputs given a dataset consisting of examples for which the labels are known, is called supervised learning.
Supervised learning requires a labeled trainning set.
In such cases, when every example is characterized by the same number of numerical features, we say that the inputs are fixed-length vectors and we call the (constant) length of the vectors the dimensionality of the data.
One major advantage of deep learning over traditional methods is the comparative grace with which modern models can handle varying-length data.
1.2.2 Models
deep learning is differentiated from classical approaches principally by the set of powerful models that it focuses on. These models consist of many successive transformations of the data that are chained together top to bottom, thus the name deep learning.
1.2.3 Objective Functions
When a model performs well on the training set but fails to generalize to unseen data, we say that it is overfitting to the training data.
1.3 Kinds of Machine Learning Problems
Supervised learning describes tasks where we are given a dataset containing both features and labels and asked to produce a model that predicts the labels when given input features.
A typical use case for supervised learning involves regression. Since we can solve simple regression problems analytically, this provides us with a solid starting point to comprehend the workings of deep learning.
The output of classfication typically lies in a given set or a range that represents possibility.
The problem of learning to predict classes that are not mutually exclusive is called multi-label classification. Auto-tagging problems are typically best described in terms of multi-label classification.
Nowadays, search engines use machine learning and behavioral models to obtain query-dependent relevance scores.
Recommender systems are another problem setting that is related to search and ranking. The problems are similar insofar as the goal is to display a set of items relevant to the user.
Specifically, sequence-to-sequence learning considers problems where both inputs and outputs consist of variable-length sequences.