Short343ef
Dołączył: 23 Gru 2010
Posty: 163
Przeczytał: 0 tematów
Ostrzeżeń: 0/5
Skąd: England
|
 Wysłany: Czw 9:15, 31 Mar 2011 Temat postu: nike air jordans Artificial Neural Ne jordan sneak |
 |
|
Artificial neural networks are the backbone of many types of AI systems. From machine learning to fuzzy logic, neural nets provide powerful processing and flexible reasoning to perform virtually any function.
Yegnanarayana, B. (2006). Artificial Neural Networks. Prentice Hall of India.
Nanotechnology & Neural Networking - In research into tiny plastic AI systems to produce better artificial brains, nanotechnology miniaturizes electrical circuits, and neural networking adds computing power.
AI neural networks are based on the same concept. Small processors are connected, with each processor given a certain threshold that it must reach prior to passing along the signal. These processors are extremely limited, and can only act on a binary signal. The result for any artificial neuron is always either yes or no, positive or negative [link widoczny dla zalogowanych], 0 or 1. If the threshold level is never reached, the processor will never pass along a signal, which causes a negative result for that path.
Neural networks, due to the complex method of information processing, are good frameworks for a specific type of artificial intelligence system. A program that requires decision-making that does not fall within clear-cut categories, works well when powered by an artificial neural net. These networks also work well with all types of structured machine learning, and but can also help machines learn independently. The simplest method of training a program, when using a neural network, is by running a learning algorithm. A learning algorithm is a separate program that gives the artificial intelligence system the information it needs to learn a task.
One key aspect of an artificial intelligence application is its ability to "learn." AI programs are able to add to algorithms and databases as necessary to improve function and increase flexibility. Fuzzy logic, decision-tree and expert systems are all types of artificial intelligence decision-making programs that use neural networks to power the necessary calculations in their complex algorithms.
Self-Improving Artificial Intelligence Programs - Artificial intelligence programs can perform many functions [link widoczny dla zalogowanych], even modifying their own code. However, they lack motivation: the drive to improve and increase knowledge.
Using Neural Networks
Neural networks in the human brain are composed of nerve cells called neurons. Each neuron is connected to thousands of others, in a complex system that provides pathways for electrical signals based on certain inputs to the neuron.
Looking for more information? Read:
Natural Neural Nets
Introduction to Machine Learning Approaches - Supervised learning [link widoczny dla zalogowanych], unsupervised learning, and reinforcement learning are methods used by artificial intelligence and explained in this introduction to machine learning.
Artificial Neural Networks
Artificial Neural Network Basics
References:
Arbib, M. (2003). The Handbook of Brain Theory and Neural Networks, Massachusetts Institute of Technology.
Fuzzy Logic Applications - The use of fuzzy logic applications in decision-making software can be complex. A simple explanation of this type of artificial intelligence (AI) programming can help.
Read on
AI Applications in Online Medical Diagnosis
A Neural Networks' Artificial Intelligence
Easy Science Fair Experiment - AI Turing Test
In AI programming, a neural network is a type of computer processor that is made to mimic the way that the human brain works. Just as it does in the brain's neural networks, the artificial neural networks used by artificial intelligence systems are composed of networks of miniature processors.
Post został pochwalony 0 razy
|
|
