ANN

ANN

What is ANN?

• ANN stands for “Artificial Neural Network”

• ANN is an algorithm, software or hardware that mimics the human brain operations.

• ANN is a part of artificial intelligence (AI).

• An artificial neural network is a combination of “structures”, “processing elements”, and “learning ability”.

How neural network works?

How neural network works in human brain?

The brain contains neurons, which are working as a switch. These neurons can change their output state based on the chemical input.

The human brain is a combination of an interconnected network of neurons. (Interconnected means one output state of any neuron is working as an input for thousand other neurons).

Human brain learns new things by activating particular neural connection and that process strengthen the interconnected network.

How neural network work in “artificial neural networks” (ANN)?

Artificial neural networks mimic the human brain operations. There are two types of learning:
• Supervised learning
• Unsupervised learning

Supervised learning: In this, ANN is trained by giving matched input and output test data. That means data is already tagged with correct answer.

For example, you trained ANN about a list of different fruit by its shape, size, and color. Now you have given a single fruit to ANN to identify. Since ANN has learned how to identify fruit name by its shape, size, and color, so it will identify the fruit name.

Unsupervised learning: In this, ANN is trained to act without any learning. Here ANN has to divide the data according to pattern, matches and differences.

For example, there are images of balls and bats. Now ANN does not know about ball and bat. Therefore, ANN will divide these images according to their pattern. Means it will divide into two groups, one for a bat and second for a ball. Here ANN divided these images without any training or pre-defined test data.

Artificial neural networks

Neural network applications

• Aircraft fault detection
• Cancer cell analysis in the healthcare domain
• Character or handwriting recognition in fraud detection (Banking sector)
• Stock price prediction
• Paraphrase identification (Two sentence’s meaning comparison)
• Speech understanding
• Spell checking
• Target tracking in the defense system
• Real-time language translation

What are the advantages of using artificial neural networks?

• Without training, ANN can produce output
• ANN learn actions and take decisions
• Can perform multiple jobs at a time
• Even in the event of failure, it still works and gives output
• Network issues do not impact immediately
• It has human-like thinking

What are the disadvantages of using artificial neural network?

• It does not explain the proofs for the output
• In some cases, it needs the training to work
• High computational load
• Takes time when input data is big

Conclusion

The neural network is a very big subject. There are lots of research going on around ANN. The neural network works well in image recognition and character recognition. The quality to learn from examples makes it very powerful. ANN contributes to the research part also.

Leave a Reply

Your email address will not be published.