neural network
- 1. Neural NetworkGuided byChapala MaharanaSubmitted By-Nilmani SinghBranch-CSERegd. No.-0601289251
- 2. Contents.Introduction to Neural Network.Neurons.Activation Function.Types of Neural Network.Learning In Neural Networks.Application of Neural Network.Advantages of Neural Network.Disadvantages Of Neural Network
- 3. Neural NetworksA method of computing, based on the interaction of multiple connected processing elements.
- 4. A powerful technique to solve many real world problems.
- 5. The ability to learn from experience in order to improve their performance.
- 6. Ability to deal with incomplete informationBasics Of Neural NetworkBiological approach to AI
- 8. Comprised of one or more layers of neurons
- 9. Several types, we’ll focus on feed-forward and feedback networksNeuronsBiologicalArtificial
- 10. Neural Network NeuronsReceives n-inputsMultiplies each input by its weightApplies activation function to the sum of resultsOutputs result
- 11. Activation FunctionsControls when unit is “active” or “inactive”Threshold function outputs 1 when input is positive and 0 otherwiseSigmoid function = 1 / (1 + e-x)
- 12. Types of Neural NetworksNeural Network types can be classified based on following attributes:Connection Type - Static (feedforward) - Dynamic (feedback) Topology - Single layer - Multilayer - Recurrent Learning Methods - Supervised - Unsupervised - Reinforcement
- 13. Classification Based On Connection TypesStatic(Feedforward)
- 14. Dynamic(Feedback)Classification Based On Topology Recurrent
- 15. Single layer
- 16. Multilayer(unit delay operator z-1implies dynamic system)
- 17. Classification Based On Learning MethodSupervisedUnsupervisedReinforcementSupervised learning Each training pattern: input + desired output
- 18. At each presentation: adapt weights
- 19. After many epochs convergence to a local minimumUnsupervised LearningNo help from the outsideNo training data, no information available on the desired outputLearning by doingUsed to pick out structure in the input:ClusteringReduction of dimensionality compressionExample: Kohonen’s Learning Law
- 21. The teacher scores the performance of the training examples
- 22. Use performance score to shuffle weights ‘randomly’
- 23. Relatively slow learning due to ‘randomness’Neural Network ApplicationsPattern recognition
- 25. Control systems & monitoring
- 26. Mobile computing
- 27. Marketing and financial applications
- 28. Forecasting – sales, market research, meteorologyAdvantages:A neural network can perform tasks that a linear program can not.
- 29. When an element of the neural network fails, it can continue without any problem by their parallel nature.
- 30. A neural network learns and does not need to be reprogrammed.
- 31. It can be implemented in any application.
- 32. It can be implemented without any problemDisadvantages:The neural network needs training to operate.
- 33. The architecture of a neural network is different from the architecture of microprocessors therefore needs to be emulated.
- 34. Requires high processing time for large neural networks.ConclusionsNeural networks provide ability to provide more human-like AITakes rough approximation and hard-coded reactions out of AI design (i.e. Rules and FSMs)Still require a lot of fine-tuning during development
- 35. ReferencesNeural Networks, Fuzzy Logic, and Genetic Algorithm ( synthesis and Application) S.Rajasekaran, G.A. VijayalakshmiPai, PHI Neuro Fuzzy and Soft Computing, J. S. R. JANG,C.T. Sun, E. Mitzutani, PHINeural Netware, a tutorial on neural networksSweetser, Penny. “Strategic Decision-Making with Neural Networks and Influence Maps”, AI Game Programming Wisdom 2, Section 7.7 (439 – 46)Russell, Stuart and Norvig, Peter. Artificial Intelligence: A Modern Approach, Section 20.5 (736 – 48)
- 36. THANK YOU
- 37. QUESTIONS ?