Kips Class 9 AI Unit 3: Introduction to Neural Network

Unit 3: Introduction to Neural Network

Question 1: Define node, weak AI system.
Answer:
Node: Artificial Neural Networks use simplified models of neurons in the brain. These neurons are called nodes.
Weak AI system: It is that intelligence that mimics intelligence but does not process it or in other words, acts intelligently without having any intelligence.

Question 2: Write about ANN.
Answer: The data enters the ANN (Artificial Neural Networks) through an input layer. The input layer communicates the data to one of the hidden layers. The hidden layer processes the data and sends the output to another hidden layer (if there are more than one hidden layer) or the output layer. The output layer provides the output.

Question 3: Mention four uses of ANN.
Answer: Four uses of ANN are
• Voice recognition system
• Language translations
• Handwriting recognition
• Automated driving vehicles

Question 4: What is the idea behind the creation of ANN?
Answer: Neural Networks are an attempt to copy the working of the interconnected cells in a human brain inside a computer. The basic idea is to create a system based on human brains, so that the computer can learn, recognise and make human-like decisions. ANN do not require any programming. Instead they learn by processing data just as human brains do.

Question 5: Explain the structure of HNN and ANN.
Answer:
• Human Neural Network (HNN): Each neuron of the brain is made up of a cell body. There are several dendrites attached to the cell body and a single axon. The dendrites carry information from outside to the cell. The axon carries the information away from the cell body.
• Artificial Neural Network (ANN): The data enters the ANN (Artificial Neural Networks) through an input layer. The input layer communicates the data to one of the hidden layers. The hidden layer processes the data and sends the output to another hidden layer (if there are more than one hidden layer) or the output layer. The output layer provides the output.

Question 6: How does Neural Network differ from Conventional computing?
Answer:
• Conventional Computer
→ A serial computer has a central processor that can address an array of memory locations where data and instructions are stored.
→ Computations are made by the processor reading instruction as well as any data that is required from memory addresses. The instructions is then executed, and the results are saved in a specified memory location, as required.
→ In a serial system, the computational steps are deterministic, logical, and sequential and the state of a given variable can be tracked from one operation to another.
• Neural Network
→ ANNs are not sequential or necessarily deterministic.
→ There are no complex central processors, instead, there are many simple ones that generally do nothing more than taking the weighed sum of their inputs from the other processors.
→ ANNs do not execute programmed instructions. They respond in parallel to the pattern of inputs presented to it.
→ There are also no separate memory addresses for storing data. Instead, information is stored in the overall activation ‘state’ of the network.
→ ‘Knowledge’ is thus represented by the network itself, which is quite literally more than the sum of its components.

Question 7: Explain how neural networks learn.
Answer: Neural networks use three ways of learning:
• Supervised Learning: In this type of learning, the networks are trained to provide correct output by using several example inputs.
• Unsupervised Learning: In this type of learning, the neural network is only provided with inputs. It doesn’t receive any information regarding the output.
• Reinforcement Learning: In this type of learning, the neural networks learn based on the feedback received by it. Positive feedback will mean that it has provided correct output and negative feedback will mean that it has made a mistake.

Question 8: Write any four applications of neural networks.
Answer: Four applications of neural networks are
• for capturing associations or discovering regularities within a set of patterns
• where the volume, number of variables or diversity of the data is very high
• if the relationships between the variable are vaguely understood
• if the relationships are difficult to describe adequately with conventional approaches

Question 9: Does ANN systems need to be programmed? Justify.
Answer: No, ANN systems do not require any programming. They instead learn by processing data as human brains do. Neural Networks are an attempt to copy the working of the interconnected cells in a human brain inside a computer. The basic idea is to create a system based on human brains, so that the computer can learn, recognise and make human-like decisions.

Question 10: Explain the importance of feedback in the learning process of ANN.
Answer: The neural networks learn through an element of feedback. The feedback tells the network if it has made a mistake or performed correctly. One of this ways of this learning involves backpropagation. In this system, the output of the neural networks is compared with the intended outputs. The weights and thresholds of the different nodes are changed to reduce the gap between the intended output and actual output. This process is repeated with every input and the system continues to learn.

Question 11: Explain the flow of information through Neural Networks.
Answer: The information that flows through the networks can take place under two distinct circumstances:
• When the system is learning or is under training
• When the system is in operation

Question 12: Explain backpropagation and feedforward network.
Answer:
• Feedforward Network: The information in the system is received by the input layer. The input layer, in turn, triggers the layers of the hidden units. The hidden units, in turn, trigger the output layer for delivering the solutions. This is one of the most common neural network designs and is known as a feedforward network.
• Backpropagation: In this system, the output of the neural networks is compared with the intended outputs. The weights and thresholds of the different nodes are changed to reduce the gap between the intended output and actual output. This process is repeated with every input and the system continues to learn. This is known as backpropagation.

Question 13: Match the components of ANN which resembles the functionality of human brains.
Answer:
Nodes → Neurons
Input Layer → Dendrites
Hidden Layer → Cell body
Output Layer → Axon