Data analytics

AI

(1) The study of human thought processes and (2) the representation and duplication of those thought process in machines.

What is word2vec, and what does it add to traditional text mining?

1) a technique used to convert words to vectors, thereby capturing their meaning, semantic similarity, and relationship with surrounding text. It grants the ability to recognize similar words.

briefly explain the three-step process of learning in ANN.

1) compute temporary outputs 2) compare outputs with desired targets 3) adjust weights & repeat process

How does CNN process text? What are word embeddings, and how do they work?

1) involves a series of filters of different sizes and shapes which convolve (roll over) the original sentence matrix to reduce it into further low dimension matrices Word embeddings-a type of word representation that allows words with similar meaning to have a similar representation. individual words are represented as real-valued vectors in a lower-dimensional space and captures inter-word semantics

List the nine steps in conducting a neural network project.

1. Define the problem 2. Gather and preprocess data 3. Design the network architecture 4. Set up training and validation 5. Initialize and train the network. 6. Validate and tune 7. Test and evaluate 8. Iterative improvement. 9. Deployment and monitoring.

What are the main steps and decision points in developing an SVM model?

1. Preprocesing the data Scrub and transform data 2. Develop the model Select kernel type (RBF is often natural choice). Determine the kernel parameters for the selected kernel type. If good results, finalize model. If not, alter it to achieve desired accuracy. 3. Deploy the model.

List and briefly describe the most commonly used ANN activation functions.

1. Sigmoid:allows the network to introduce non-linearity into the model, which allows the neural network to learn more complex decision boundaries. Rectified Linear unit (ReLU) is an activation function that introduces the property of nonlinearity to a deep learning model and solves the vanishing gradients issue. Hyperbolic tangent (tanh): The tanh function is similar to the sigmoid function but with a range between -1 and 1. It is commonly used in neural networks for binary classification and regression problems. Softmax: The softmax function is commonly used in multiclass classification problems where the output is a probability distribution over multiple classes. It ensures that the sum of the outputs is equal to 1. Linear: The linear function simply returns the input value without any modification. It is used in regression problems where the output can have any value. Leaky ReLU: The leaky ReLU function is a modified version of ReLU that returns a small positive value for negative inputs instead of 0. This helps to prevent the "dying ReLU" problem where a neuron can become permanently inactive.

How does backpropagation learning work?

1.Initialize weights with random values 2.Read in the input vector and the desired output 3.Compute the actual output via the calculations 4.Compute the error. Change the weights by working backward

How does a feedforward multilayer perceptron-type deep network work?

A feedforward neural network or multilayer perceptrons (MLPs) is an artificial neural network wherein connections between the units do not form a cycle. Information flows through the function being evaluated from x, through the intermediate computations used to define f, and finally to the output y.Sep 8, 2017

What is the "kernel trick"? How is it used in SVM?

A method for converting a linear classifier algorithm into a nonlinear one by using a nonlinear function to map the original observations into a high-dimensional space. The kernel trick is used to map the input data points into a higher-dimensional space where the separation between the two classes becomes easier.

What is MLP, and how does it work?

A multilayer perceptron (MLP) is a fully connected feed-forward artificial neural network with at least three layers (input, output, and at least one hidden layer). works only in the forward direction. All nodes are fully connected to the network. Each node passes its value to the coming node only in the forward direction. The MLP neural network uses a Backpropagation algorithm to increase the accuracy of the training model.

artificial brain

A people-made machine that is designed to be as intelligent, creative, and self-aware as humans.

The major benefits of AI are

AI has the ability to complete certain tasks much faster than humans. The consistency of the completed AI work can be much better than that of humans. AI machines do not make mistakes. AI systems allow for continuous improvement projects. AI can be used for predictive analysis via its capability of pattern recognition. AI can manage delays and blockages in business processes. AI machines do not stop to rest or sleep. AI machines can work autonomously or be assistants to humans. The functionalities of AI machines are ever increasing. AI machines can learn and improve their performance. AI machines can work in environments that are hazardous to people. AI machines can facilitate innovations by human (i.e., support research and development [R&D]). No emotional barriers interfere with AI work. AI excels in fraud detection and in security facilitations. AI improves industrial operations. AI optimizes knowledge work. AI increases speed and enables scale. AI helps with the integration and consolidating of business operations. AI applications can reduce risk. AI can free employees to work on more complex and productive jobs. AI improves customer care. AI can solve difficult problems that previously were unsolved (Kharpal, 2017). AI increases collaboration and speeds up learning

What are the major capabilities of human intelligence? Which are superior to that of AI machines?

AI is best suited for handling repetitive, data-driven tasks and making data-driven decisions. However, human skills such as creativity, critical thinking, emotional intelligence, and complex problem-solving still need to be more valuable and easily replicated by AI.

How intelligence is AI

AI machines have demonstrated superiority over humans in playing complex games such as chest, jeopardy, and Go whose top players were beaten by a computer using the well-known program. Is difficult comparing AI to human intelligence as it is a multidimensional situation.

What is the significance of AlexNet? Draw and describe its architecture.

AlexNet is a pioneering convolutional neural network (CNN) used primarily for image recognition and classification tasks. consists of 5 convolutional layers, 3 max-pooling layers, 2 normalization layers, 2 fully connected layers, and 1 softmax layer. 2. Each convolutional layer consists of convolutional filters and a nonlinear activation function ReLU.

What are the commonalities and differences between biological and artificial neural networks?

An artificial neural network's processing was sequential and centralized. In contrast, a biological neural network processes information in parallel and distributive. The artificial neural network is of a much smaller size than the biological neural network. In contrast, the biological neural network is large in size.

Define deep learning.

Attempts to mimic how the human brain works. Deep learning uses artificial neural technology and plays a major role in dealing with complex application that regular machine learning and other AI technologies cannot handle.

Define CPU, NVIDIA, CUDA, and deep learning, and comment on the relationship between them.

CPU-- The CPU is the primary component of a computer that performs most of the processing inside a computer. It interprets instructions and carries out arithmetic, logic, and control functions specified by the computer's instructions. NVIDIA--a multinational technology company renowned for designing graphics processing units (GPUs) for gaming, professional visualization, data centers, and automotive markets. CUDA--a parallel computing platform and application programming interface (API) model created by NVIDIA. It allows developers to utilize the parallel computing power of NVIDIA GPUs for general-purpose processing tasks, not just graphics rendering. In summary, NVIDIA GPUs, powered by CUDA, play a vital role in accelerating deep learning tasks by providing the computational power necessary for training complex neural network models on large datasets.

What is cognitive computing, and how does it differ from other computing paradigms?

Cognitive Computing uses algorithms and data analysis to recognize patterns and make decisions, while Artificial Intelligence uses machine learning algorithms. Cognitive computing aims to emulate human-like intelligence and cognitive processes, while traditional computing focuses on executing predefined instructions. Cog Comp systems are data driven where traditional computing relies on static algorithms Better problem solving and contextual awareness

How does cognitive computing differ from ordinary AI techniques?

Cognitive computing uses the same technologies as AI but with the addition of self-learning programs such as data mining and pattern recognition.

List and briefly describe three different types of LSTM applications.

Common LSTM applications include sentiment analysis, language modeling, speech recognition, and video analysis.

What is convolution function in CNN and how does it work?

Convolution is a mathematical operation that allows the merging of two sets of information. In the case of CNN, convolution is applied to the input data to filter the information and produce a feature map. This filter is also called a kernel, or feature detector, and its dimensions can be, for example, 3x3.

Examples of Augmented intelligence

Cybercrime fighting. For example, AI can identify forthcoming attacks and suggest solutions. E-commerce decisions. AI marketing tools can make testing results 100 times faster, and adapt the layout and response functions of a Web site to users. Machines also make recommendations, and marketers can accept or reject them. High-frequency stock market trading. This process can be done either completely autonomously or in some cases with human control and calibration.

Despite the short tenure of deep learning implementation, why do you think there are several different computing frameworks for it?

Deep learning (DL) frameworks offer building blocks for designing, training, and validating deep neural networks through a high-level programming interface.Dec 1, 2022

What is meant by "deep" in deep neural networks? Compare deep neural networks to shallow neural networks.

Deep neural networks are similar to other machine learning methods as they use sophisticated mathematical algorithms to learn from data similar to the way that humans learn. Shallow networks typically have only one or two hidden layers, while deep networks can have dozens or even hundreds of layers.

Intelligence agents and their capabilities

Definition- an autonomous, relatively small computer software program that observes and acts upon changes in its environment by running specific tasks autonomously.

How does Hopfield ANN architecture work? To what type of problems can it be applied?

Each neuron has an inverting and a non-inverting output. Being fully connected, the output of each neuron is an input to all other neurons but not the self. The Hopfield network is commonly used for auto-association and optimization tasks.

What are the common application areas for SVM? Search the Internet to identify popular application areas and specific SVM software tools used in those applications.

Face detection, IMage classification, text catagorization, generalized predictive control.

What is GoogleNet? How does it work?

GoogLeNet is a convolutional neural network that is 22 layers deep. The network trained on ImageNet classifies images into 1000 object categories, such as keyboard, mouse, pencil, and many animals.

What is GPU? How does it relate to deep neural networks?

Graphics processing units (GPUs), originally developed for accelerating graphics processing, can dramatically speed up computational processes for deep learning. They are an essential part of a modern artificial intelligence infrastructure, and new GPUs have been developed and optimized specifically for deep learning.

What are the typical use cases for cognitive analytics?

Healthcare, analyze data, fraud detection, increases efficiency, and cybersecurity.

What is ImageNet and how does it relate to deep learning?

ImageNet is an image database organized according to the WordNet hierarchy (currently only the nouns), in which each node of the hierarchy is depicted by hundreds and thousands of images. The project has been instrumental in advancing computer vision and deep learning research.

What is overfitting in ANN learning? How does it happen, and how can it be mitigated?

In Overfitting, the model tries to learn too many details in the training data along with the noise from the training data. As a result, the model performance is very poor on unseen or test datasets. You can prevent overfitting by diversifying and scaling your training data set

What is Keras, and how is it different from the other frameworks?

Keras is like a user-friendly tool for building and training neural networks, which are computer systems inspired by the human brain. In essence, Keras stands out for its simplicity, ease of use, and high-level abstraction, making it a preferred choice for many deep learning practitioners.

What is the LSTM network, and how does it differ from RNNs?

LSTM is a type of RNN with higher memory power to remember the outputs of each node for a more extended period to produce the outcome for the next node efficiently. LSTM networks combat the RNN's vanishing gradients or long-term dependence issue.

Limitations of AI machines

Lack human touch and feel Lack attention to non-task surroundings Can lead people to rely on AI machines too much (e.g., people may stop to think on their own) Can be programmed to create destruction (see discussion in Chapter 14) Can cause many people to lose their jobs (see Chapter 14) Can start to think by themselves, causing significant damage (see Chapter 14)

What is machine learning? How can it be used in business?

Machine learning classifies as "teaching computers to learn from examples and large amounts of data. Improves customer loyalty and retention Hiring the right people Automating finance Detecting fraud Providing predictive maintenance Providing retail shelf analysis Making other prediction like in medicine to investments.

Describe machine translation of languages. Why it is important in business?

Machine translation uses computer programs to translate words and sentences from one language to another. Machine translation helps to lower or eliminate the language barrier in communication. Individuals quickly obtain a translation of text.

3 levels of AI: Augmented intelligence

Most of the existing AI applications, which are between assisted and autonomous, are referred to as augmented intelligence (or intelligence augmentation). Their technology can augment computer tasks to extend human cognitive abilities

What are the most popular neural network architectures?

Multilayer perceptron with backpropagation), associative memory, recurrent networks, Kohonen's self-organizing feature maps, and Hopfield networks.

What is NLP? What are its two major formats?

NLP is a technology that gives users the ability to communicate with a computer in their active language. Natural language understanding that investigates methods of enabling computers to comprehend instructions or queries provided in ordinary English or other human languages. Natural language generation that strives to have computers produce ordinary spoken language so that people can understand the computers more easily

What is an ANN?

Neural computing refers to a pattern-recognition methodology for machine learning. The resulting model from neural computing is often called an artificial neural network (ANN).

The use of AI has been driven by the following forces:

People's interest in smart machines and artificial brains The low cost of AI applications versus the high cost of manual labor (doing the same work) The desire of large tech companies to capture competitive advantage and market share of the AI market and their willingness to invest billions of dollars in AI The pressure on management to increase productivity and speed The availability of quality data contributing to the progress of AI The increasing functionalities and reduced cost of computers in general The development of new technologies, particularly cloud computing

Major goals of AI

Perceive and properly react to changes in the environment that influence specific business process and operations. Introduce creativity in business processes and decision making.

Describe the different types of neural network software available today.

Perceptron Feed Forward Neural Network Multilayer Perceptron Convolutional Neural Network Radial Basis Functional Neural Network Recurrent Neural Network LSTM - Long Short-Term Memory Sequence to Sequence Models Modular Neural Network

How can one measure the intelligence level of a vacuum cleaner?

Provided that measuring the intelligence of AI is complex and multifaceted involving a combination of quantitative metrics. Some are: Task Performance: One of the most straightforward ways to measure AI intelligence is by evaluating its performance on specific tasks. Generalization: Intelligent AI systems should be able to generalize their knowledge and skills beyond the specific tasks they were trained on Learning Efficiency: Assessing how quickly and effectively an AI system can learn from new data or experiences can be indicative of its intelligence. Robustness and Resilience: Intelligent AI systems should be robust and resilient to variations, noise, and adversarial attacks. Creativity and Novelty: Some AI systems are designed to exhibit creativity or generate novel solutions to problems.

What types of problems are solved with Kohonen SOM ANN architecture?

Provides a way to represent multidimensional data in much lower dimensional spaces, usually one or two dimensions. Learns to supervise data without supervision

How do you determine the optimal kernel type and kernel parameters?

RBF is generally the best kernel type while selecting the parameters that fit the problem the best.

What is the significance of "context," "sequence," and "memory" in RNN?

RNNs have a memory of past inputs, which allows them to capture information about the context of the input sequence. This makes them useful for tasks such as language modeling, where the meaning of a word depends on the context in which it appears.Aug 25, 2023

What is RNN? How does it differ from CNN?

Recurrent neural networks (RNNs) are a class of neural network that are helpful in modeling sequence data. CNNs are feedforward neural networks that use filters and pooling layers, whereas RNNs feed results back into the network.

What is augmented reality?

Refers to the integration of digital information with the user environment in real time.

Describe the summation and activation functions in MLP-type ANN architecture.

Summation functions-all features are multiplied by their weights and bias are summed up and then applied over an Activation function.

List and briefly define the characteristics of different deep learning frameworks.

TensorFlow: TensorFlow is an open-source deep learning framework developed by Google Brain Team. It provides a flexible architecture with high-level APIs for building and training deep learning models. TensorFlow supports both symbolic and imperative programming, allowing for easy experimentation and deployment. It offers extensive community support, a rich ecosystem of tools and libraries, and deployment options for various platforms, including mobile and edge devices. PyTorch: PyTorch is an open-source deep learning framework developed by Facebook's AI Research lab (FAIR). It emphasizes dynamic computation graphs, making it more intuitive and Pythonic compared to TensorFlow's static graph approach. PyTorch provides a simple and flexible API, making it easier for researchers and developers to prototype and experiment with new ideas. It has gained popularity for its ease of use, strong GPU acceleration support, and state-of-the-art performance in research and production settings. Keras: Keras is an open-source deep learning library written in Python that serves as an interface for TensorFlow, Theano, and Microsoft Cognitive Toolkit (CNTK). It provides a user-friendly and modular API for building neural networks, allowing for rapid prototyping and experimentation. Keras supports both convolutional and recurrent networks, as well as combinations of the two. It is known for its simplicity, ease of use, and beginner-friendly design, making it a popular choice for beginners and experts alike. MXNet: MXNet is an open-source deep learning framework developed by Apache Software Foundation. It offers scalability and efficiency, with support for distributed computing across multiple GPUs and devices. MXNet provides a clean and concise API with both imperative and symbolic programming paradigms. It is known for its high performan

What is cognitive computing?

The application of knowledge derived from cognitive science and computer science theories in order to simulate the human thought processes so that computers can exhibit and/or support decision-making and problem-solving capabilities.

List and briefly define the key attributes of cognitive computing.

The cognitive computing process uses a combination of artificial intelligence, machine learning, neural networks, sentiment analysis, natural language processing, and contextual awareness to solve daily problems just like human beings.

Compared to ANN, what are the advantages of SVM?

The development of SVM involved sound statistical learning theory first and then implementation and experiments; whereas ANN did not. The solutions of SVMs are also global and unique where ANN is not. SVM also does not suffer from overfitting like ANNs.

What are the advantages and disadvantages of SVM?

The main advantage of SVM is its ability to handle high-dimensional data, while a disadvantage is its sensitivity to noise and outliers. The main advantage of SVM is its high accuracy, while the absence of negative data for training is a disadvantage.

What is the meaning of "maximum-margin hyperplanes"? Why are they important in SVM?

The maximum-margin hyperplanes are used to optimally separate the output classes from each other in the training data. These are crucial to avoid effects of random noise and overfitting. Its use also leads to the smaller VC dimension and reduces the # of potential classifiers.

Limitations of the Turing test.

The test focuses solely on the ability to mimic human-like behavior and communication, rather than on the underlying intelligence or consciousness of the machine

3 levels of AI decisions: Autonomous AI

These systems are in the realm of the strong AI but in a very narrow domain. Eventually, a computer will take over many tasks, automating them completely. Machines act as experts and have absolute decision-making power. Pure robo-advisors (Chapter 12) are examples of such machines. Autonomous vehicles and robots that can fix themselves are also good examples.

What are knowledge systems?

These systems, which are presented in Chapter 12, are computer programs that store knowledge, which their applications use to generate expert advice and/or perform problem solving. Knowledge-based expert systems also help people to verify information and make certain types of automated routine decisions.

What types of business problems can be solved with ANN?

They can classify information, cluster data, or predict outcomes. ANN's can be used for a range of tasks. These include analyzing data, transcribing speech into text, powering facial recognition software, or predicting the weather

For what type of applications can CNN be used?

They find applications in image recognition, object detection, facial recognition, medical image analysis, self-driving cars, and more.

3 levels of AI Decisions: Assisted intelligence

This is equivalent mostly to weak AI, which works only in narrow domains. It requires clearly defined inputs and outputs. Examples are some monitoring systems and low-level virtual personal assistants (Chapter 12). Such systems and assistants are used in our vehicles for giving us alerts. Similar systems can be used in many healthcare applications (e.g., monitoring, diagnosing).

Explain what the terms cognitive analytics and cognitive search mean.

To put it simply, a cognitive analytics system searches through the data that exists in its knowledge base to find solutions that make sense for the questions posed. Cognitive analytics can be thought of as analytics with human-like intelligence. a search engine technology that uses artificial intelligence (AI) to quickly find relevant and accurate search results for various types of queries

How do we measure AI intelligence

Turing test- A test to determine whether computers are intelligence when a human interviewer questions a human and a machine is unable to determine which is which. // To pass the Turing Test, a computer needs to be able to understand a human language (NLP), to possess human intelligence (e.g., have a knowledge base), to reason using its stored knowledge, and to be able to learn from its experiences (machine learning).

Which strategy is better: more layers or more neurons

Usually, you will get more of a performance boost from adding more layers than from adding more neurons in each layer.

What is IBM Watson and what is its significance to the world of computing?

Watson was created as a question answering (QA) computing system that IBM built to apply advanced natural language processing, information retrieval, knowledge representation, automated reasoning, and machine learning technologies to the field of open domain question answering. Watson demonstrated that machines could understand and interact in a natural language, question-and-answer format and learn from their mistakes.

What are some of the design parameters for developing a neural network?

Weights and biases

Explain the function of weights in ANN.

Weights in ANNs are analogous to the synapses in biological neural networks. They are the parameters that are adjusted during the training process to minimize the difference between the actual output and the target output.

How do SVM work?

Works by mapping data to a high-dimensional feature space so that data points can be categorized, even when the data are not otherwise linearly separable. A separator between the categories is found, then the data are transformed in such a way that the separator could be drawn as a hyperplane.

Draw and explain the functioning of a typical recurrent neural network unit.

a deep learning model that is trained to process and convert a sequential data input into a specific sequential data output.

How does a single artificial neuron (i.e., PE) work?

a function that receives one or more inputs, applies weights to these inputs, and sums them to produce an output.

How do Google's Neural Machine Translation and Microsoft Skype Translator work?

by first breaking the input text into smaller units, such as words or phrases. The neural network then learns to map these units to their corresponding units in the target language.

Intelligence

can be considered an umbrella term and is usually measured by an IQ test. It is composed of reasoning, learning, logic, problem-solving ability, perception, and linguistic ability.

List and briefly explain five use cases for IBM Watson.

helps your underwriters, claims processors, customer service agents and actuaries find answers and insights from insurance documents, customer and public data faster

What is pooling in CNN? How does it work?

involves sliding a two-dimensional filter over each channel of feature map and summarizing the features lying within the region covered by the filter

Prepare a list of applications of intelligent agents.

searching for products, getting recommendations, determining products' pricing, planning marketing, improving computer security, managing auctions, facilitating payments, and improving inventory management.

Define robotics and explain its importance for manufacturing and transportation.

the intersection of science, engineering and technology that produces machines, called robots, that replicate or substitute for human actions. Used to automate various tasks, such as loading and unloading cargo, inspecting vehicles, and performing maintenance task. Some manufacturers use robotics to automate repetitive, menial tasks such as material handling and assembly.

How does Watson work?

uses natural language processing, a technology that analyzes human speech for meaning and syntax.