Lecture 11 neural network principles

Jan 7, 20190 likes335 views

Vajira Thambawita

Basic of neural networks has been introduced here as a supportive knowledge to understand basic of robotics.

Neural Network
• The artificial neural network (ANN), often simply called neural
network (NN), is a processing model loosely derived from biological
neurons.
• Neural networks are often used for classification problems or decision
making problems that do not have a simple or straightforward
algorithmic solution.
• The beauty of a neural network is its ability to learn an input to
output mapping from a set of training cases without explicit
programming, and then being able to generalize this mapping to
cases not seen previously.
• We concentrate on the topics relevant to mobile robots.

Neural Network Principles
• Individual artificial neuron
Sigmoidal output
function

Feed-Forward Networks
• Fully connected feed-forward network (differ significantly from
feedback networks)

Three-layer NN
• For most practical applications, a single hidden layer is sufficient.
• Input layer (for example input from robot sensors)
• Hidden layer (connected to input and output layer)
• Output layer (for example output to robot actuators)
• In the standard three-layer network, the input layer is usually
simplified in the way that the input values are directly taken as
neuron activation. No activation function is called for input neurons.

Three-layer NN
The remaining questions for our standard three-layer NN type are:
1. How many neurons to use in each layer?
The number of neurons in the input and output layer are determined by the
application. (three PSD sensors as input, two motors as output
 the network should have three input neurons and two output neurons ).
2. Which connections should be made between layer i and layer i + 1?
We simply connect every output from layer i to every input at layer i + 1. This
is called a “fully connected” neural network.
3. How are the weights determined?
The standard method is supervised learning, for example through error
backpropagation. The same task is repeatedly run by the NN and the
outcome judged by a supervisor.

Three-layer NN
• Errors made by the network are backpropagated from the output
layer via the hidden layer to the input layer, amending the weights of
each connection.
• Evolutionary algorithms provide another method for determining the
weights of a neural network. A genetic algorithm can be used to
evolve an optimal set of neuron weights.

Example
• The experimental setup for an NN that should drive a mobile robot collision-free
through a maze (for example left-wall following) with constant speed.
• Three sensor inputs + two motor outputs + chose six hidden neurons (3+6+2)

Example
• Let us calculate the output of an NN for a simpler case with 2 + 4 + 1
neurons.

Backpropagation
• A large number of different techniques exist for learning in neural
networks.
• These include supervised and unsupervised techniques.
• Classification networks  a supervised off-line technique, can be
used to identify a certain situation from the network input and
produce a corresponding output signal.

1. The document discusses various image filtering techniques, including correlation filtering, convolution, averaging filters, and Gaussian filters. 2. Gaussian filters are commonly used for smoothing images as they remove high-frequency components while maintaining edges. The scale parameter σ controls the amount of smoothing. 3. Median filters can reduce noise in images by selecting the median value in a local neighborhood, unlike mean filters which are susceptible to outliers.

Artificial neural networkIshaneeSharma

The document discusses artificial neural networks (ANNs). It describes ANNs as computing systems composed of interconnected processing elements that mimic the human brain. ANNs can solve complex problems in parallel and are fault tolerant. The key components of an ANN are the input, hidden and output layers. Feedforward and feedback networks are described. Backpropagation is used to train ANNs by adjusting weights and biases based on error. Training can be supervised, unsupervised or reinforced learning. Patterns and batch modes of training are also outlined.

Artificial neural networkmustafa aadel

Neural networks can be biological models of the brain or artificial models created through software and hardware. The human brain consists of interconnected neurons that transmit signals through connections called synapses. Artificial neural networks aim to mimic this structure using simple processing units called nodes that are connected by weighted links. A feed-forward neural network passes information in one direction from input to output nodes through hidden layers. Backpropagation is a common supervised learning method that uses gradient descent to minimize error by calculating error terms and adjusting weights between layers in the network backwards from output to input. Neural networks have been applied successfully to problems like speech recognition, character recognition, and autonomous vehicle navigation.

Artificial Neural Networks: Introduction, Neural Network representation, Appr...BMS Institute of Technology and Management

neuralnetwork.pptxSherinRappai1

This document discusses artificial neural networks. It defines neural networks as computational models inspired by the human brain that are used for tasks like classification, clustering, and pattern recognition. The key points are: - Neural networks contain interconnected artificial neurons that can perform complex computations. They are inspired by biological neurons in the brain. - Common neural network types are feedforward networks, where data flows from input to output, and recurrent networks, which contain feedback loops. - Neural networks are trained using algorithms like backpropagation that minimize error by adjusting synaptic weights between neurons. - Neural networks have many applications including voice recognition, image recognition, robotics and more due to their ability to learn from large amounts of data.

neuralnetwork.pptxSherinRappai

This document discusses artificial neural networks. It defines neural networks as computational models inspired by the human brain that are used for tasks like classification, clustering, and pattern recognition. The key points are: - Neural networks contain interconnected artificial neurons that can perform complex computations. They are inspired by biological neurons in the brain. - Common neural network types are feedforward networks, where data flows from input to output, and recurrent networks, which contain feedback loops. - Neural networks are trained using algorithms like backpropagation that minimize error by adjusting synaptic weights between neurons. - Neural networks have various applications including voice recognition, image recognition, and robotics due to their ability to learn from large amounts of data.

2011 0480.neural-networksParneet Kaur

The document discusses artificial neural networks and backpropagation. It provides background on neural networks, including their biological inspiration from the human brain. It describes the basic components of artificial neurons and how they are connected in networks. It explains feedforward neural networks and discusses limitations of single-layer perceptrons. The document then introduces multi-layer feedforward networks and the backpropagation algorithm, which allows training of hidden layers by propagating error backwards. It provides details on calculating error terms and updating weights in backpropagation training.

Feed forward back propogation algorithm .pptxneelamsanjeevkumar

The document provides an introduction to the back-propagation algorithm, which is commonly used to train artificial neural networks. It discusses how back-propagation calculates the gradient of a loss function with respect to the network's weights in order to minimize the loss through methods like gradient descent. The document outlines the history of neural networks and perceptrons, describes the limitations of single-layer networks, and explains how back-propagation allows multi-layer networks to learn complex patterns through error propagation during training.

Introduction to Artificial Intelligence_ Lec 9Dalal2Ali

employed to cover the tampering traces of a tampered image. Image tamperingrapellisrikanth

1.Introduction to Artificial Neural Networks.pptxsalahidddin

Neural networkKRISH na TimeTraveller

Neural networks are mathematical models inspired by biological neural networks. They are useful for pattern recognition and data classification through a learning process of adjusting synaptic connections between neurons. A neural network maps input nodes to output nodes through an arbitrary number of hidden nodes. It is trained by presenting examples to adjust weights using methods like backpropagation to minimize error between actual and predicted outputs. Neural networks have advantages like noise tolerance and not requiring assumptions about data distributions. They have applications in finance, marketing, and other fields, though designing optimal network topology can be challenging.

Artificial Neural Networks Artificial Neural NetworksMajdDassan

NeuralVaibhav Shah

This document provides an overview of neural networks. It discusses how neural networks were inspired by biological neural systems and attempt to model their massive parallelism and distributed representations. It covers the perceptron algorithm for learning basic neural networks and the development of backpropagation for learning in multi-layer networks. The document discusses concepts like hidden units, representational power of neural networks, and successful applications of neural networks.

Jyduydufyuyf8yfiyfiyfifiyfiyviyviyfiugiuy8f7dd64d4yrsxyfhgdhfjhvjhvtestankitkush

Unit 2 ml.pptxPradeeshSAI

- The document discusses multi-layer perceptrons (MLPs), a type of artificial neural network. MLPs have multiple layers of nodes and can classify non-linearly separable data using backpropagation. - It describes the basic components and working of perceptrons, the simplest type of neural network, and how they led to the development of MLPs. MLPs use backpropagation to calculate error gradients and update weights between layers. - Various concepts are explained like activation functions, forward and backward propagation, biases, and error functions used for training MLPs. Applications mentioned include speech recognition, image recognition and machine translation.

Introduction to Neural networks (under graduate course) Lecture 9 of 9Randa Elanwar

Basics of Artificial Neural Network Subham Preetam

This document provides an overview of artificial neural networks. It discusses the biological neuron model that inspired artificial neural networks. The key components of an artificial neuron are inputs, weights, summation, and an activation function. Neural networks have an interconnected architecture with layers of nodes. Learning involves modifying the weights through algorithms like backpropagation to minimize error. Neural networks can perform supervised or unsupervised learning. Their advantages include handling complex nonlinear problems, learning from data, and adapting to new situations.

19_Learning.pptgnans Kgnanshek

This document discusses neural networks and their learning capabilities. It describes how neural networks are composed of simple interconnected elements that can learn patterns from examples through training. Perceptrons are introduced as single-layer neural networks that can learn linearly separable functions through a simple learning rule. Multi-layer networks are shown to have greater learning capabilities than perceptrons using an algorithm called backpropagation that propagates errors backward through the network to update weights. Applications of neural networks include pattern recognition, control problems, and time series prediction tasks.

Artificial neural network by arpit_sharmaEr. Arpit Sharma

This document provides an introduction to artificial neural networks. It discusses biological neurons and how artificial neurons are modeled. The key components of a neural network including the network architecture, learning approaches, and the backpropagation algorithm for supervised learning are described. Applications and advantages of neural networks are also mentioned. Neural networks are modeled after the human brain and learn by modifying connection weights between nodes based on examples.

UNIT 5-ANN.pptSivam Chinna

Neural networks are inspired by biological neurons and are used to learn relationships in data. The document defines an artificial neural network as a large number of interconnected processing elements called neurons that learn from examples. It outlines the key components of artificial neurons including weights, inputs, summation, and activation functions. Examples of neural network architectures include single-layer perceptrons, multi-layer perceptrons, convolutional neural networks, and recurrent neural networks. Common applications of neural networks include pattern recognition, data classification, and processing sequences.

neuralnetwork.pptxKarthik Rohan

Neural networks are programs that mimic the human brain by learning from large amounts of data. They use simulated neurons that are connected together to form networks, similar to the human nervous system. Neural networks learn by adjusting the strengths of connections between neurons, and can be used to perform tasks like pattern recognition or prediction. Common neural network training algorithms include gradient descent and backpropagation, which help minimize errors by adjusting connection weights.

02 Fundamental Concepts of ANNTamer Ahmed Farrag, PhD

This document provides an overview of neural networks and fuzzy systems. It outlines a course on the topic, which is divided into two parts: neural networks and fuzzy systems. For neural networks, it covers fundamental concepts of artificial neural networks including single and multi-layer feedforward networks, feedback networks, and unsupervised learning. It also discusses the biological neuron, typical neural network architectures, learning techniques such as backpropagation, and applications of neural networks. Popular activation functions like sigmoid, tanh, and ReLU are also explained.

Artificial Neural Networks presentationsmigob991

Competitive Neural Network Artificial Intelligence21118057

Neuralnetwork 101222074552-phpapp02Deepu Gupta

Neural networks are a new method of programming computers that are good at pattern recognition. They are inspired by the human brain and are composed of interconnected processing elements called neurons. Neural networks learn by example through adjusting synaptic connections between neurons. They can be trained to perform tasks like pattern recognition and classification. There are different types of neural networks including feedforward and feedback networks. Training involves adjusting weights to minimize error through algorithms like backpropagation. Neural networks are used in applications like data analysis, forecasting, and medical diagnosis.

Machine learning by using python lesson 2 Neural Networks By Professor Lili S...Professor Lili Saghafi

This document provides an overview of lesson 2 of a machine learning course using Python. It discusses neural networks and their biological inspiration. It then explains how artificial neural networks work, including the basic neuron structure and how signals are received and transmitted. Finally, it introduces implementing simple neural networks in Python using NumPy for efficient data structures.

Lecture 4 principles of parallel algorithm design updatedVajira Thambawita

Lecture 3 parallel programming platformsVajira Thambawita

Parallel platforms can be organized in various ways, from an ideal parallel random access machine (PRAM) to more conventional architectures. PRAMs allow concurrent access to shared memory and can be divided into subclasses based on how simultaneous memory accesses are handled. Physical parallel computers use interconnection networks to provide communication between processing elements and memory. These networks include bus-based, crossbar, multistage, and various topologies like meshes and hypercubes. Maintaining cache coherence across multiple processors is important and can be achieved using invalidate protocols, directories, and snooping.

More Related Content

Similar to Lecture 11 neural network principles (20)

Introduction to Artificial Intelligence_ Lec 9Dalal2Ali

employed to cover the tampering traces of a tampered image. Image tamperingrapellisrikanth

1.Introduction to Artificial Neural Networks.pptxsalahidddin

Neural networkKRISH na TimeTraveller

Artificial Neural Networks Artificial Neural NetworksMajdDassan

NeuralVaibhav Shah

Jyduydufyuyf8yfiyfiyfifiyfiyviyviyfiugiuy8f7dd64d4yrsxyfhgdhfjhvjhvtestankitkush

Unit 2 ml.pptxPradeeshSAI

Introduction to Neural networks (under graduate course) Lecture 9 of 9Randa Elanwar

Basics of Artificial Neural Network Subham Preetam

19_Learning.pptgnans Kgnanshek

Artificial neural network by arpit_sharmaEr. Arpit Sharma

UNIT 5-ANN.pptSivam Chinna

neuralnetwork.pptxKarthik Rohan

02 Fundamental Concepts of ANNTamer Ahmed Farrag, PhD

Artificial Neural Networks presentationsmigob991

Competitive Neural Network Artificial Intelligence21118057

Neuralnetwork 101222074552-phpapp02Deepu Gupta

Machine learning by using python lesson 2 Neural Networks By Professor Lili S...Professor Lili Saghafi

Introduction to Artificial Intelligence_ Lec 9Dalal2Ali

employed to cover the tampering traces of a tampered image. Image tamperingrapellisrikanth

1.Introduction to Artificial Neural Networks.pptxsalahidddin

Neural networkKRISH na TimeTraveller

Artificial Neural Networks Artificial Neural NetworksMajdDassan

NeuralVaibhav Shah

Jyduydufyuyf8yfiyfiyfifiyfiyviyviyfiugiuy8f7dd64d4yrsxyfhgdhfjhvjhvtestankitkush

Unit 2 ml.pptxPradeeshSAI

Introduction to Neural networks (under graduate course) Lecture 9 of 9Randa Elanwar

Basics of Artificial Neural Network Subham Preetam

19_Learning.pptgnans Kgnanshek

Artificial neural network by arpit_sharmaEr. Arpit Sharma

UNIT 5-ANN.pptSivam Chinna

neuralnetwork.pptxKarthik Rohan

02 Fundamental Concepts of ANNTamer Ahmed Farrag, PhD

Artificial Neural Networks presentationsmigob991

Competitive Neural Network Artificial Intelligence21118057

Neuralnetwork 101222074552-phpapp02Deepu Gupta

Machine learning by using python lesson 2 Neural Networks By Professor Lili S...Professor Lili Saghafi

More from Vajira Thambawita (20)

Lecture 4 principles of parallel algorithm design updatedVajira Thambawita

Lecture 3 parallel programming platformsVajira Thambawita

Lecture 2 more about parallel computingVajira Thambawita

Lecture 1 introduction to parallel and distributed computingVajira Thambawita

Lecture 12 localization and navigationVajira Thambawita

Localization and navigation are important tasks for mobile robots. Localization involves determining a robot's position and orientation, which can be done using global positioning systems outdoors or local sensor networks indoors. Navigation involves planning a path to reach a goal destination. Common navigation algorithms include Dijkstra's algorithm, A* algorithm, potential field method, wandering standpoint algorithm, and DistBug algorithm. Each algorithm has different requirements and approaches to planning paths between a starting point and goal.

Lecture 10 mobile robot designVajira Thambawita

Lecture 09 controlVajira Thambawita

On-off control is the simplest method of feedback control where the motor power is either switched fully on or off depending on whether the actual speed is higher or lower than the desired speed. A PID controller is a more advanced control method that uses proportional, integral and derivative terms to provide smoother control compared to on-off control and help reduce steady-state error. PID control is almost an industry standard approach for feedback-based motor speed regulation.

Lecture 08 robots and controllersVajira Thambawita

Sensors and actuators are important components for robots. Sensors can be analog or digital and include sensors for position, orientation, distance, light, and more. The right sensor must match the application needs. Actuators allow robots to move and interact with their environment. Common actuators include DC motors, stepper motors, and servos, which can be controlled through techniques like pulse-width modulation. Together, sensors and actuators enable robots to perceive and interact with the world.

Lecture 07 more about picVajira Thambawita

Lecture 06 pic programming in cVajira Thambawita

Lecture 05 pic io port programmingVajira Thambawita

Lecture 04 branch call and time delayVajira Thambawita

Lecture 03 basics of picVajira Thambawita

Lecture 02 mechatronics systemsVajira Thambawita

Mechatronics is the synergistic combination of mechanical, electrical, and computer engineering with an emphasis on integrated design. It has applications across many scales, from micro-electromechanical systems to large transportation systems like high-speed trains. Some key applications discussed in the document include CNC machining, automobiles using technologies like brake-by-wire, smart home appliances, prosthetics, pacemakers and defibrillators, unmanned aerial vehicles, and robots for space exploration, military, sanitation, and other uses. Mechatronics allows the development of advanced, integrated systems for improved performance, safety, efficiency and user experience.

Lecture 1 - Introduction to embedded system and RoboticsVajira Thambawita

Lec 09 - Registers and CountersVajira Thambawita

Registers are groups of flip-flops that store binary information, while counters are a special type of register that sequences through a set of states. A register consists of flip-flops and gates, and can store multiple bits. Counters increment or decrement their state in response to clock pulses. There are two main types: ripple counters where flip-flops trigger each other, and synchronous counters where all flip-flops change on a clock pulse.

Lec 08 - DESIGN PROCEDUREVajira Thambawita

Design procedures or methodologies specify hardware that will implement the desired behaviour. The design of a clocked sequential circuit starts from a set of specifications and culminates in a logic diagram or a list of Boolean functions from which the logic diagram can be obtained. More informations: https://sites.google.com/view/vajira-thambawita/leaning-materials/slides

Lec 07 - ANALYSIS OF CLOCKED SEQUENTIAL CIRCUITSVajira Thambawita

Lec 06 - Synchronous Sequential LogicVajira Thambawita

Lec 05 - Combinational LogicVajira Thambawita

Lecture 4 principles of parallel algorithm design updatedVajira Thambawita

Lecture 3 parallel programming platformsVajira Thambawita

Lecture 2 more about parallel computingVajira Thambawita

Lecture 1 introduction to parallel and distributed computingVajira Thambawita

Lecture 12 localization and navigationVajira Thambawita

Lecture 10 mobile robot designVajira Thambawita

Lecture 09 controlVajira Thambawita

Lecture 08 robots and controllersVajira Thambawita

Lecture 07 more about picVajira Thambawita

Lecture 06 pic programming in cVajira Thambawita

Lecture 05 pic io port programmingVajira Thambawita

Lecture 04 branch call and time delayVajira Thambawita

Lecture 03 basics of picVajira Thambawita

Lecture 02 mechatronics systemsVajira Thambawita

Lecture 1 - Introduction to embedded system and RoboticsVajira Thambawita

Lec 09 - Registers and CountersVajira Thambawita

Lec 08 - DESIGN PROCEDUREVajira Thambawita

Lec 07 - ANALYSIS OF CLOCKED SEQUENTIAL CIRCUITSVajira Thambawita

Lec 06 - Synchronous Sequential LogicVajira Thambawita

Lec 05 - Combinational LogicVajira Thambawita

Recently uploaded (20)

Low Vison introduction from Aligarh Muslim UniversityAligarh Muslim University, Aligarh, Uttar Pradesh, India

"Orthoptera: Grasshoppers, Crickets, and Katydids pptxArshad Shaikh

Philosophical Basis of Curriculum DesigningAnkit Choudhary

The philosophical basis of curriculum refers to the foundational beliefs and values that shape the goals, content, structure, and methods of education. Major educational philosophies—idealism, realism, pragmatism, and existentialism—guide how knowledge is selected, organized, and delivered to learners. In the digital age, understanding these philosophies helps educators and content creators design curriculum materials that are purposeful, learner-centred, and adaptable for online environments. By aligning educational content with philosophical principles and presenting it through interactive and multimedia formats.

Research Handbook On Environment And Investment Law Kate Milesmucomousamir

Sri Guru Arjun Dev Ji .Balvir Singh

Optical and non optical devices used in low visionAligarh Muslim University, Aligarh, Uttar Pradesh, India

Unit 1 Kali NetHunter is the official Kali Linux penetration testing platform...ChatanBawankar

Multicultural approach in education - B.Edprathimagowda443

EVALUATION AND MANAGEMENT OF OPEN FRACTUREBipulBorthakur

Quiz-E-Mataram (Under 20 Quiz Set) .pptxSouptikUkil

New-Beginnings-Cities-and-States.pdf/7th class social/4th chapterFor online c...Sandeep Swamy

New Beginnings: Cities and States This presentation explores the Second Urbanisation of India, examining the rise of janapadas and mah janapadas as crucial developments in India's early history. by sandeep swamy The Second Urbanisation Urban Revival The emergence of new cities after the decline of the Indus Valley Civilization. Historical Timeline Occurred approximately between 600-200 BCE in the Gangetic plains. New Settlements Formation of organized urban centers with political and economic significance. Janapadas: Early States Definition Territorial units with distinct cultural and political identities. Significance Formation Evolved from tribal settlements into more organized political entities. Marked the transition from nomadic to settled agricultural communities. Magadha: The Powerful Kingdom Strategic Location Situated in modern-day Bihar with natural defenses of hills and rivers. Access to iron ore deposits gave military advantage. Capital Cities R jagr#iha (Rajgir) served as the initial capital. Later shifted to P t#aliputra (modern Patna). Ruins of a major structure at R jagr#iha, the early capital of Magadha.

Unit 4 Reverse Engineering Tools Functionalities & Use-Cases.pdfChatanBawankar

Writing Research Papers: Guidance for Research CommunityRishi Bankim Chandra Evening College, Naihati, North 24 Parganas, West Bengal, India

Understanding-the-Weather.pdf/7th class/social/ 2nd chapter/Samyans Academy n...Sandeep Swamy

QUIZ-O-FORCE PRELIMINARY ANSWER SLIDE.pptxSourav Kr Podder

Patent Law in Bangladesh Addressing Challenges in Pharmaceutical Innovation a...Ibrahim Tareq

How to create and manage blogs in odoo 18Celine George

KNN,Weighted KNN,Nearest Centroid Classifier,Locally Weighted RegressionGlobal Academy of Technology

LDMMIA About me 2025 Edition 3 College VolumeLDM & Mia eStudios

Updated About Me. Used for former college assignments. Make sure to catch our weekly updates. Updates are done Thursday to Fridays or its a holiday/event weekend. Thanks again, Readers, Guest Students, and Loyalz/teams. This profile is older. I started at the beginning of my HQ journey online. It was recommended by AI. AI was very selective but fits my ecourse style. I am media flexible depending on the course platform. More information below. AI Overview: “LDMMIA Reiki Yoga refers to a specific program of free online workshops focused on integrating Reiki energy healing techniques with yoga practices. These workshops are led by Leslie M. Moore, also known as LDMMIA, and are designed for all levels, from beginners to those seeking to review their practice. The sessions explore various themes like "Matrix," "Alice in Wonderland," and "Goddess," focusing on self-discovery, inner healing, and shifting personal realities.”

Decision Tree-ID3,C4.5,CART,Regression TreeGlobal Academy of Technology

Low Vison introduction from Aligarh Muslim UniversityAligarh Muslim University, Aligarh, Uttar Pradesh, India

"Orthoptera: Grasshoppers, Crickets, and Katydids pptxArshad Shaikh

Philosophical Basis of Curriculum DesigningAnkit Choudhary

Research Handbook On Environment And Investment Law Kate Milesmucomousamir

Sri Guru Arjun Dev Ji .Balvir Singh

Optical and non optical devices used in low visionAligarh Muslim University, Aligarh, Uttar Pradesh, India

Unit 1 Kali NetHunter is the official Kali Linux penetration testing platform...ChatanBawankar

Multicultural approach in education - B.Edprathimagowda443

EVALUATION AND MANAGEMENT OF OPEN FRACTUREBipulBorthakur

Quiz-E-Mataram (Under 20 Quiz Set) .pptxSouptikUkil

New-Beginnings-Cities-and-States.pdf/7th class social/4th chapterFor online c...Sandeep Swamy

Unit 4 Reverse Engineering Tools Functionalities & Use-Cases.pdfChatanBawankar

Writing Research Papers: Guidance for Research CommunityRishi Bankim Chandra Evening College, Naihati, North 24 Parganas, West Bengal, India

Understanding-the-Weather.pdf/7th class/social/ 2nd chapter/Samyans Academy n...Sandeep Swamy

QUIZ-O-FORCE PRELIMINARY ANSWER SLIDE.pptxSourav Kr Podder

Patent Law in Bangladesh Addressing Challenges in Pharmaceutical Innovation a...Ibrahim Tareq

How to create and manage blogs in odoo 18Celine George

KNN,Weighted KNN,Nearest Centroid Classifier,Locally Weighted RegressionGlobal Academy of Technology

LDMMIA About me 2025 Edition 3 College VolumeLDM & Mia eStudios

Decision Tree-ID3,C4.5,CART,Regression TreeGlobal Academy of Technology

Lecture 11 neural network principles

1. Neural Network

2. Neural Network • The artificial neural network (ANN), often simply called neural network (NN), is a processing model loosely derived from biological neurons. • Neural networks are often used for classification problems or decision making problems that do not have a simple or straightforward algorithmic solution. • The beauty of a neural network is its ability to learn an input to output mapping from a set of training cases without explicit programming, and then being able to generalize this mapping to cases not seen previously. • We concentrate on the topics relevant to mobile robots.

3. Neural Network Principles • Individual artificial neuron Sigmoidal output function

4. Feed-Forward Networks • Fully connected feed-forward network (differ significantly from feedback networks)

5. Three-layer NN • For most practical applications, a single hidden layer is sufficient. • Input layer (for example input from robot sensors) • Hidden layer (connected to input and output layer) • Output layer (for example output to robot actuators) • In the standard three-layer network, the input layer is usually simplified in the way that the input values are directly taken as neuron activation. No activation function is called for input neurons.

6. Three-layer NN The remaining questions for our standard three-layer NN type are: 1. How many neurons to use in each layer? The number of neurons in the input and output layer are determined by the application. (three PSD sensors as input, two motors as output  the network should have three input neurons and two output neurons ). 2. Which connections should be made between layer i and layer i + 1? We simply connect every output from layer i to every input at layer i + 1. This is called a “fully connected” neural network. 3. How are the weights determined? The standard method is supervised learning, for example through error backpropagation. The same task is repeatedly run by the NN and the outcome judged by a supervisor.

7. Three-layer NN

8. Three-layer NN • Errors made by the network are backpropagated from the output layer via the hidden layer to the input layer, amending the weights of each connection. • Evolutionary algorithms provide another method for determining the weights of a neural network. A genetic algorithm can be used to evolve an optimal set of neuron weights.

9. Example • The experimental setup for an NN that should drive a mobile robot collision-free through a maze (for example left-wall following) with constant speed. • Three sensor inputs + two motor outputs + chose six hidden neurons (3+6+2)

10. Example • Let us calculate the output of an NN for a simpler case with 2 + 4 + 1 neurons.

11. Example

12. Example Activation value Output values

13. Backpropagation • A large number of different techniques exist for learning in neural networks. • These include supervised and unsupervised techniques. • Classification networks  a supervised off-line technique, can be used to identify a certain situation from the network input and produce a corresponding output signal.

14. Backpropagation

15. Bias neurons

16. summary

Lecture 11 neural network principles

Recommended

More Related Content

Similar to Lecture 11 neural network principles (20)

More from Vajira Thambawita (20)

Recently uploaded (20)

Lecture 11 neural network principles