Artificial Intelligence

When it comes to making complex judgement calls, computers can’t replace people. But with artificial intelligence, computers could be trained to think like humans do. Artificial intelligence allows computers to learn from experience, recognize patterns in large amounts of complex data and make complex decisions based on human knowledge and reasoning skills. Artificial intelligence has become an important field of study with a wide spread of applications in fields ranging from medicine to agriculture.

Expert Systems

Two of the most important and most used branches of AI are neural networks and expert systems.

An expert system can solve real-world problems using human knowledge and following human reasoning skills. Knowledge and thinking processes of experts are collected and encoded into a knowledge base. From that point on, the expert system could replace or assist the human experts in making complex decisions by integrating all the knowledge it has in its knowledge base.

Neural Networks

Illustration of Neural NetworkThis diagram represents an artificial neural network. A neural network is made of nodes arranged in different patterns representing the "intelligence" of the network. The line thickness indicates the strength of the connections.









The most important application of neural networks is in pattern recognition. Humans, through neurons in their brains, learn how to read human writing, recognize a bad apple from a good one or identify their children from a set of kids. Neural networks allow computers to use the same principles that neurons in the brains use to recognize and classify different patterns. So in a way, neural networks are a digital representation (although very simplified) of our brains. They are made of artificial neurons, connected by weights, which are indicative of the strengths of the connections. The neurons are arranged in layers, and depending on the complexity of the application, there could be a few of them or a very large number of them (hundreds or thousands). Iterative propagation of input from one layer of neurons to the next (training) is what enables the neural network to learn from experience.

Unlike humans, when a neural is fully trained, it can classify and identify patterns in massive amounts of complex data. It could do this at high speeds that can not be duplicated by humans.

Real-World Applications of Artificial Intelligence

Intelligent control is beneficial in many real world applications because it is good at solving complex problems. The flexibility inherent in AI techniques, makes the technology adaptable to fields as diverse as agriculture, business, and literature.

UGA scientists have used artificial intelligence in many different ways: monitor and adjust the climate in greenhouses and poultry production houses help forecast weather and predict crop development determine when vegetables are ripe identify molecules by their "chemical fingerprints" candle eggs to determine which have cracks and other defects.