book

Artificial Neural Networks in Food Processing

Name: Artificial Neural Networks in Food Processing
Author: Mohamed Tarek Khadir
ISBN: 9783110646139

by Mohamed Tarek Khadir

January 2021

Intermediate to advanced

200 pages

6h 22m

English

De Gruyter

Read now

Unlock full access

Acknowledgement
Introduction
1 Biological inspiration and single artificial neurons
1.1 Introduction and biological inspiration1.2 Brief history of ANN developmentORIGINSTHE QUIET YEARSINTEREST REVIVAL1.3 Artificial neural network types1.4 Simple formal neurons1.4.1 Mc Culloch–Pitt neuron1.4.2 The perceptron1.5 Hebb learning rule (1949)1.5.1 The Hebb learning rule applied to a sign-type neuron1.5.2 Hebb learning rule used for a Perceptron type neuron1.6 Perceptron learning rule1.7 Classification and linear separability1.7.1 Classification1.7.2 Discriminant functions1.7.3 Linear separability1.7.4 Example: Implementation of the AND function using a single neuron1.8 Linear combiner and the LMS algorithm1.9 Gradient descent optimisation1.9.1 Stability and learning rate1.9.2 Generalisation to a multi-dimensional system1.10 LMS learning algorithm1.10.1 The Adaline1.10.2 The Delta learning rule1.10.3 Derivation of the Delta learning rule1.10.4 Derivative evaluation for the Delta learning rule
2 Artificial neural networks for food processes: a survey
2.1 Machine perception2.1.1 Electronic nose2.1.2 Electronic tongue2.1.3 Machine vision2.2 Spectroscopic data interpretation2.3 Food microbiology and food fermentation2.4 Other applications2.5 Overview of ANN applications and perspectives classified by food types2.5.1 Dairy and dairy products2.5.2 Meat and fish2.5.3 Fruits and vegetables2.5.4 Cereals and grains2.5.5 Tea2.5.6 Alcoholic beverages2.5.7 Oil2.5.8 Other products
3 Multi-layered perceptron
3.1 Artificial neural networks: principles and training algorithms3.2 Regressor type and network topologies3.3 Multi-layer perceptron topology3.4 Activation function3.5 Cybenko theorem3.6 Network training3.6.1 Supervised learning3.6.2 Unsupervised learning3.6.3 Back-propagation learning3.6.4 Error function derivatives3.6.5 Update process3.7 Exclusive OR (XOR) classification using MLP3.7.1 Analytical solution3.7.2 Adaptive solution3.8 Enhancing MLP’s training3.8.1 Inertia moment approach3.8.2 Adaptive learning step3.8.3 Weight update using new entries3.9 Practical considerations when designing MLPs3.9.1 Data normalisation3.9.2 Informative training and validation data for MLPs3.9.3 Training procedures3.9.4 Topological optimisation of MLPs3.10 Case study: Application of MLP for wine classification3.10.1 Wine dataset used3.10.2 Creation and training a MLP for wine classification3.10.3 Validation and results assessment
4 Radial basis function networks
4.1 Radial basis function network topology4.2 Radial basis function network’s training algorithm4.2.1 K-Means algorithm4.2.2 Radial basis function curse of dimensionality4.3 Classification of simple examples using RBF: the XOR problem4.4 Case study: classification of oil origin using spectral information and RBFs4.4.1 Database and available data4.4.2 Training and validation of the RBF network using random centring
5 Self-organising feature maps or Kohonen maps
5.1 Self-organising maps architectures5.2 Self-organising maps training5.2.1 Competition phase5.2.2 Adaptation phase5.2.3 Initialisation and parameterising of the Kohonen map5.2.4 Visualisation of Kohonen maps results5.3 Case study: clustering of substrate methane production using Kohonen maps5.3.1 Clustering substrate by their biogas production capabilities5.3.2 Available data and experimental results5.3.3 Two stages Kohonen map approach for substrate methane production clustering5.3.4 Conclusions
6 Deep artificial neural networks
6.1 Introduction6.1.1 Definition of deep learning6.1.2 Principles of deep learning6.2 Architectures of deep neural networks6.2.1 Recurrent neural network (RNN)6.2.2 Auto-Encoder (AE)6.2.3 Deep Belief Network (DBN)6.2.4 Boltzmann machine6.2.5 Recurrent Neural Network (RNN)6.2.6 Convolutional Neural Network (CNN)6.3 Bolzmann Machine (BM)6.4 Recurrent neural networks: Long Short Term Memory Networks (LSTM)6.4.1 Forward pass6.4.2 Backward pass6.4.3 Update process6.5 Stack Denoising Auto-Encoder (SDAE)6.5.1 Auto-Encoder (AE)6.5.2 Denoising Auto-Encoder (DAE)6.5.3 Stacking Denoising Auto-Encoder (SDAE)6.6 Convolutional Neural Network (CNN)6.6.1 Training steps of a convolutional neural network6.6.2 Choice of the hyperparameters6.6.3 CNNs architectures6.7 Convolutional Stacked Auto-Encoder Neural Networks (CSAE)6.8 Case study: Deep neural networks for fruits recognition and classification6.8.1 Fruit classification and recognition using convolutional neural networks6.8.2 Fruit classification and recognition using stacked convolutional auto-encoder6.8.3 Fruit classification and recognition using stacked convolutional denoising auto-encoder
7 Overview of model predictive control theory and applications in food science using ANN
7.1 ANN and model predictive control in food process applications7.1.1 Strategy of using ANNs in a MPC approach7.1.2 Overview for ANN based MPC in food processes7.1.3 Baking applications7.1.4 Drying applications7.1.5 Fermentation/brewing applications7.1.6 Thermal/pressure applications7.1.7 Dairy applications7.1.8 Other food process applications7.2 Principles of linear model predictive control7.2.1 Introduction7.2.2 Origins of model predictive control7.3 Linear model predictive control algorithms7.3.1 First-order predictive functional control7.3.2 Generalised predictive control7.4 Nonlinear model predictive control7.5 Case study: neural networks and model predictive control for pasteurisation temperature control in a plate heat exchanger7.5.1 Milk pasteurisation process7.5.2 Justification and rationale for using MPC7.5.3 Modelling of the pasteurization plant using ANN7.5.4 Neural predictive controller design
Subject Index

Content preview from Artificial Neural Networks in Food Processing

5 Self-organising feature maps or Kohonen maps

Self-organizing maps commonly designated by SOFM (for Self-organizing Feature Maps, or SOM for Self-organizing Maps) were introduced by T. Kohonen in 1981 [2], inspired by how human brain topographic maps work, where near points in the human body are represented by groups of nearby neurons in the brain. These maps are not uniform, and are the most sensitive surface of the human body and are represented by a cluster containing the largest number of neurons. From an IT point of view, we can translate this property as follows: suppose we have a set of data that we want to classify. We are looking for a mode of representation such as objects neighbours are classified in the same class or in neighbouring ...

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.

Read now

Unlock full access

More than 5,000 organizations count on O’Reilly

O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.

Julian F.

Head of Cybersecurity

I wanted to learn C and C++, but it didn't click for me until I picked up an O'Reilly book. When I went on the O’Reilly platform, I was astonished to find all the books there, plus live events and sandboxes so you could play around with the technology.

Addison B.

Field Engineer

I’ve been on the O’Reilly platform for more than eight years. I use a couple of learning platforms, but I'm on O'Reilly more than anybody else. When you're there, you start learning. I'm never disappointed.

Amir M.

Data Platform Tech Lead

I'm always learning. So when I got on to O'Reilly, I was like a kid in a candy store. There are playlists. There are answers. There's on-demand training. It's worth its weight in gold, in terms of what it allows me to do.

Mark W.

Embedded Software Engineer

The Hidden Cost of AI-Assisted Creativity

Publisher Resources

ISBN: 9783110646139

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills

Artificial Neural Networks in Food Processing

by Mohamed Tarek Khadir

5 Self-organising feature maps or Kohonen maps

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.