Many problems in AI (including in reasoning, planning, learning, perception, and robotics) require the agent to operate with incomplete or uncertain information. AI researchers have devised a number of tools to solve these problems using methods from probability theory and economics.[86]
Bayesian networks[87] are a very general tool that can be used for many problems, including reasoning (using the Bayesian inference algorithm),[g][89] learning (using the expectation-maximization algorithm),[h][91] planning (using decision networks)[92] and perception (using dynamic Bayesian networks).[93]
Probabilistic algorithms can also be used for filtering, prediction, smoothing and finding explanations for streams of data, helping perception systems to analyze processes that occur over time (e.g., hidden Markov models or Kalman filters).[93]
Precise mathematical tools have been developed that analyze how an agent can make choices and plan, using decision theory, decision analysis,[94] and information value theory.[95] These tools include models such as Markov decision processes,[96] dynamic decision networks,[93] game theory and mechanism design.[97]
Classifiers and statistical learning methods
The simplest AI applications can be divided into two types: classifiers (e.g. "if shiny then diamond"), on one hand, and controllers (e.g. "if diamond then pick up"), on the other hand. Classifiers[98] are functions that use pattern matching to determine the closest match. They can be fine-tuned based on chosen examples using supervised learning. Each pattern (also called an "observation") is labeled with a certain predefined class. All the observations combined with their class labels are known as a data set. When a new observation is received, that observation is classified based on previous experience.[46]
There are many kinds of classifiers in use. The decision tree is the simplest and most widely used symbolic machine learning algorithm.[99] K-nearest neighbor algorithm was the most widely used analogical AI until the mid-1990s, and Kernel methods such as the support vector machine (SVM) displaced k-nearest neighbor in the 1990s.[100] The naive Bayes classifier is reportedly the "most widely used learner"[101] at Google, due in part to its scalability.[102] Neural networks are also used as classifiers.[103]
Artificial neural networks
Artificial neural networks[103] were inspired by the design of the human brain: a simple "neuron" N accepts input from other neurons, each of which, when activated (or "fired"), casts a weighted "vote" for or against whether neuron N should itself activate. In practice, the input "neurons" are a list of numbers, the "weights" are a matrix, the next layer is the dot product (i.e., several weighted sums) scaled by an increasing function, such as the logistic function. "The resemblance to real neural cells and structures is superficial", according to Russell and Norvig.[104][i]
Learning algorithms for neural networks use local search to choose the weights that will get the right output for each input during training. The most common training technique is the backpropagation algorithm.[105] Neural networks learn to model complex relationships between inputs and outputs and find patterns in data. In theory, a neural network can learn any function.[106]
In feedforward neural networks the signal passes in only one direction.[107] Recurrent neural networks feed the output signal back into the input, which allows short-term memories of previous input events. Long short term memory is the most successful network architecture for recurrent networks.[108] Perceptrons[109] use only a single layer of neurons, deep learning[110] uses multiple layers. Convolutional neural networks strengthen the connection between neurons that are "close" to each other – this is especially important in image processing, where a local set of neurons must identify an "edge" before the network can identify an object.[111]
No comments:
Post a Comment