Complete the following assignment in one MS word document:
Chapter 5 –discussion questions #1-4 (page # 308) & exercise 6 (page # 310).
1. What is an artificial neural network and for what types of problems can it be used?
2. Compare artificial and biological neural networks. What aspects of biological networks are not mimicked by artificial ones? What aspects are similar?
3. What are the most common ANN architectures? For what types of problems can they be used?
4. ANN can be used for both supervised and unsupervised learning. Explain how they learn in a supervised mode and in an unsupervised mode.
Exercise 6: Go to Google Scholar (scholar.google.com). Conduct a search to find two papers written in the last five years that compare and contrast multiple machine-learning methods for a given problem domain. Observe commonalities and differences among their findings and prepare a report to summarize your understanding.
Chapter 6– discussion questions #1-5 (page # 383) & exercise 4 (page # 384).
1. What is deep learning? What can deep learning do that traditional machine-learning methods cannot?
2. List and briefly explain different learning paradigms methods in AI.
3. What is representation learning, and how does it relate to machine learning and deep learning?
4. List and briefly describe the most commonly used ANN activation functions.
5. What is MLP, and how does it work? Explain the function of summation and activation weights in MLP-type ANN.
Exercise 4: Cognitive computing has become a popular term to define and characterize the extent of the ability of machines/computers to show “intelligent” behavior. Thanks to IBM Watson and its success on Jeopardy!, cognitive computing and cognitive analytics are now part of many real world intelligent systems. In this exercise, identify at least three application cases where cognitive computing was used to solve complex real-world problems. Summarize your findings in a professionally organized report.
When submitting work, be sure to include an APA cover page and include at least two APA formatted references (and APA in-text citations) to support the work this week.
All work must be original (not copied from any source).
-
analytics-data-science-artificial-intelligence-systems-for-decision-support-11th-edition.pdf
SAMPLE ANSWER
1. What is an artificial neural network and for what types of problems can it be used?
Introduction
Artificial neural networks are a type of machine learning that uses artificial neurons to model the behavior of biological neurons. There are several different types of artificial neural networks, each with its own strengths and weaknesses. The most popular ones are Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs) and Long Short Term Memory (LSTM). The following article will give you all necessary info on these three types so that you can pick up which one suits your needs best!
1. What is an artificial neural network and for what types of problems can it be used?
A neural network is a collection of interconnected processing units (neurons) that interact with each other and with their surroundings. The neurons are connected to one another using weighted values called synaptic weights and biases, which can be adjusted to learn from new data. This allows a neural network to mimic the way in which the brain learns from experience.
Neural networks have been used for many years in machine learning applications such as image recognition, speech recognition, natural language processing and information retrieval systems for text databases such as PubMed or Google Knowledge Vault
2. What is activation function and how do they work?
Activation functions are the rules that determine how your network will respond to different inputs. They can be used to train a network in one direction or another, depending on what kind of activation function you choose to use.
For example, if you want your neural network to learn only positive values and ignore negative ones, then it’s best not to use an activation function that gives equal weighting for both positive and negative numbers (such as: 0/1). Instead, try using an activation function that gives greater weighting for positive values than negative ones (such as: 1/0). This will ensure that your neural network learns only neutral data points—which are exactly what we want!
3. How to pick the right number of layers and nodes in a neural network?
The number of layers and nodes in your neural network depends on the problem you are trying to solve. If you are only using a single layer, then there will be no need for any neurons outside this layer. However, if you want to train your model with many additional layers, then it might be necessary for each neuron in each additional layer to have access to all previous layers’ state information (e.g., training data).
This means that when we talk about how many neurons we should use in our model, we need another factor: the number of inputs per neuron (or equivalently: the number of output neurons per input). This is called normalization and it’s very important because without normalization everything else goes wrong! How much normalization do I need? Well…that depends on what kind of problems we’re solving with our models!
4. What is the difference between a Deep Learning model and a regular Machine Learning model?
A neural network is a model that learns from data. It is often used to learn patterns in data, such as identifying images or speech. A deep learning model is one that has multiple layers of neurons, each layer with fewer connections than the previous layer (e.g., two layers of neurons connected by 20 connections each). In addition to having more complex mathematical operations performed on them, these models are also able to extract more information from their training datasets than standard ML models do—for example, they can find objects in images when humans would fail at it! However, this comes at a cost: it takes longer for these models train because there are so many variables involved; additionally there are many ways in which you could choose how much detail/complexity should go into your final solution (for example: Do I want my model trained only on photos? Or do I want it trained on text?)
5. In what type of situations will you choose to use an unsupervised Neural Network?
Unsupervised learning is a type of machine learning that uses unlabeled data for training. Unsupervised neural networks are useful in situations where you have lots of unlabeled data, or if you want to explore the data and find hidden patterns, like clusters in images.
For example, if I had a lot of photos from my friends who visited me over the past year and I wanted to find out what kind of activities they did together, then I might use an unsupervised network with some pre-built models for this task – these would help me identify groups based on similarities between their pictures (e.g., “the team at work”).
6. How does backpropagation work in neural networks? And what are its benefits compared to other approaches used for training Artificial Neural Networks (ANNs)?
The process of backpropagation is essentially a supervised learning algorithm. It compares the output of the network to its target output and uses this error to adjust weights in order to make future predictions more accurate.
The first step of backpropagation involves finding all weights that have been used by previous time steps, as well as their magnitude (how much they contribute) and their direction (whether they point towards positive or negative values). Then, these parameters are updated based on an error function that takes into account both their magnitude and direction:
A lot of stuff about machine learning
Machine learning is a field of computer science that deals with the creation and analysis of algorithms that learn from data. In machine learning, we teach computers to find patterns in large amounts of data and make decisions based on those patterns.
Neural networks are a type of machine learning algorithm, which uses artificial neurons (the basic unit for processing information) to process information about how your brain works. Neural networks are used for many applications such as image recognition, speech parsing, handwriting recognition and even language translation!
Backpropagation training helps you train your neural network by adjusting each weight according to its output value during training so that you can improve accuracy over time without having to manually adjust each weight individually (this is called supervised learning).
Conclusion
So, you can see that there are a lot of things to take into account when it comes to neural networks. As we’ve seen, they’re incredibly powerful and have a lot of potential, but they also have their downsides. If you’re thinking about using them in your business or product development, make sure that you understand what type of problem they’re best suited for and how they work before making any decisions.