Sie sind auf Seite 1von 14

SEMINAR COMPARISON BETWEEN NEURAL NETWORKS AND COMPUTER

PRESENTED BY PANKAJ SHARMA

What is a Neural Network?


Neural network is collection of highly interconnected processing elements called neurons. Neural networks are inspired by the way biological nervous system such as brain process information. Neural networks learn by example.

Basic structure of Brain

Specifications of brain
Number of Neurons
Number of synapses/neuron Total number of synapses Operations/s/neuron Total operations Brain weight

100 billion
1000 100000 billion 100 10000 trillion 1.5 kg

Characteristics of ANN
Parallel computation: program is executed simultaneously. Robustness: insensitive to partial inputs. Self learning: algorithm is created itself. Fault tolerance: output is not affected if one or two neuron are damaged.

Basic model of an ANN


x1*w1 + x2*w2 + > th

Learning in ANN

Precautions before learning process


1. Decide the function of neural network to be performed. 2. Make a complete set of input and output patterns. 3. Determine the number of layers in network and number of nodes per layer. 4. Select the appropriate threshold value. 5. Determine the algorithm termination criteria.

Applications of neural networks

1. 2. 3. 4. 5.

In signal processing. Image data processing. Communication systems. Intelligent control. Optimization techniques.

Computer: essential characteristics


The problem to be solved should have some mathematical solution. All the steps for solving any problem should be explicitly specified. The data should have a precise format. There should not be any memory crash.

Steps in computing
1. Fetch an instruction from memory. 2. Fetch any data required by the instruction from memory. 3. Execute the instruction. 4. Store the result in memory. 5. go to step 1

Computer: advantage and limitation


1. Advantage: Fast arithmetic. Do precisely what the programmer wants to do. Makes life easier. Limitation : Not fault tolerance. Cannot solve a problem without a mathematical solution. No parallelism. Cannot adapt according to circumstances.

2.

Comparison between computer and neural network


1. Problem solving approach : in most of cases computer uses top down approach in problem solving. Whereas neural network uses bottom up approach in problem solving. Way of functioning : computers function logically with a set of rules and calculations whereas neural networks can function via images , pictures and concepts. Self programming : computers need some algorithm for problem solving , whereas neural network learns by itself. Speed : speed of computer depends on various aspects of processor , whereas neural network use chips for different applications.

2.

3.

4.

if (IsTired()) { GoToSleep();} else { StayAwake();}

A revised version:
if (IsTired()) { if (!IsInClass() && !WorkingOnProject()) { GoToSleep(); } else { if (IsInClass()) { Stay Awake(); } else { if (WorkingOnProject()) { if (AssignmentIsDueTomorrow()) { if (AssignmentIsCompleted()) { GoToSleep(); } else { StayAwake(); } } else { GoToSleep(); } } else { StayAwake(); } }} else { StayAwake();}

Das könnte Ihnen auch gefallen