editors posts
Artificial intelligence: Related terms

Artificial intelligence: Related terms

Does Artificial intelligence stimulate human intelligence?

In 1956 Dartmouth conference

  • AI is about teaching machines to learn, adapt and think while enabling them with cognitive capabilities, the complex series of layer of algorithm, using data, can solve problems for better than human. These algorithms can be used to find deep patterns.
  • AI also deals with simulation of intelligence behavior in computers. These processes include learning, reasoning, planning, problem solving etc.
  • It is used to predict, automate, and optimize tasks that humans have historically done, such as speech and facial recognition, decision making, and translation.

There are three main categories of AI:

  • Artificial Narrow Intelligence (ANI)
  • Artificial General Intelligence (AGI)
  • Artificial Super Intelligence (ASI)

ANI is considered “weak” AI, whereas the other two types are classified as “strong” AI. Weak AI is defined by its ability to complete a very specific task, like winning a chess game or identifying a specific individual in a series of photos. As we move into stronger forms of AI, like AGI and ASI, the incorporation of more human behaviors becomes more prominent, such as the ability to interpret tone and emotion. Only narrow intelligence is available now.

Computer vision:

Computer vision is a field of artificial intelligence (AI) that enables computers and systems to derive meaningful information from digital images, video and other visual inputs – and take actions or make recommendations based on that information. If AI enables computers to think, computer vision enables them to observe and understand.

Two essential technologies are used to accomplish this: a type of machine learning called deep learning and a convolutional neural network (CNN).

Here are a few examples of established computer vision tasks:

  • (Image classification, Object detection , Object tracking  ,Content-based image retrieval )

 Machine learning:

Machine learning is a branch of (AI) and computer science which focuses on the use of data and algorithms to imitate the way that humans learn, gradually improving its accuracy.

  • ML is the science of enabling computers to act without programming and learn from patterns derived from data.
  • ML is the fastest growing approach to achieve AI.
  • (NLP) Neural language processing is a form of AI that gives the ability to machine process the human language by computer program – spam detection
  • Deep learning is a subset of machine learning that can be thought of as automation of prediction analysis.

learning system of a machine learning algorithm into three main parts:

  1. Decision Process
  2. Error Function
  3. Model Optimization Process

 Machine learning classifiers fall into three primary categories.

A classifier in machine learning is an algorithm that automatically orders or categorizes data into one or more of a set of “classes”.

Machine learning is divided into two types: supervised machine learning and unsupervised machine learning.

  1. Supervised machine learning

Supervised learning is defined by its use of labeled datasets to train algorithms that to classify data or predict outcomes accurately. This training dataset includes inputs and correct outputs, which allow the model to learn over time. Supervised learning can be separated into two types of classification and regression.

  • Classification uses an algorithm to accurately assign test data into specific categories.
  • Regression is used to understand the relationship between dependent and independent variables.

Algorithms:

Support Vector Machines: Support Vector Machine (SVM) is another most powerful algorithm. It strong regularization and can be leveraged both for classification or regression challenges.

K-nearest neighbor,

KNN is a simple algorithm that stored all the available cases and classifies the new data or case based on the similarity measure.

KNN is used for both classification and regression.

It is mostly used to classifies a data point on how its neighbors are classified.

  • ‘ K ‘ in KNN is a parameter that refers to the number of nearest neighbor to inlude in the majority of voting process, if k=5 then count the nearest five neighbors and classify the object there.
  • Choosing the right value of ‘K’ is a process of parameter tunning .
  • Choosing “smaller value” of ‘K’ can be noisy and have higher influence on result.
  • Choosing “large value” of ‘K’ can produce bias.
  • Keep the value of ‘K’ odd.
  • K= sqrt(N), N is the no of sample in training data set.
  • Another way to choose ‘K’ is cross validation.
  • Choose validation dataset from training set and test on different values of ‘ K’ .
  • Minimize the validation error.
  • Training error: “average error” the result from using the machine learning method on training set 80% of data, easy to calculate.
  • Testing error: “average error” that result from using machine learning method to predict response on new observation

Training error is not good estimate of Test error: with model flexibility training error, decrease however “zero training error” is overfitting which generalized poorly.

  • KNN algorithm is defined accordingly a distance metric between two data points. A popular one is Euclidian distance method

Convolutional neural networks (CNN), used in computer vision and image classification application. It can detect features and patterns with and image enabling tasks:  object or voice detection, image recognition or processing and machine translation.

Applications: Image tagging, Visual Search, Recommender engines

Recurrent neural networks (RNN), it is a type of neural network where the output from previous step is fed as input to the current step. The main feature or ‘RNN’ is (hidden state), which remember some information about the sequence.

applications: text processing, grammar checks, Text to speech processing, Image tagger, Translation etc.)

2. Unsupervised learning

It uses ML algorithm to analyze and cluster unlabeled dataset

does not have labeled outputs, its goal is to infer the natural structure present within a set of data points.

The two main tasks of unsupervised learning are:

  • clustering data into groups by similarity.

Algorithms: k- mean clustering, Hierarchical clustering.

  • reducing dimensionality: compress the data while maintaining its structure and usefulness.

Algorithms: principal component analysis (PCA), singular value decomposition (SVD)

 

speech synthesis: the process of generating spoken language by machine based on written input.

 Deep learning

  • A subset of machine learning that can be thought of as automation of prediction analysis.
  • Predictive analysis I an approach to find insights to fine anything that any data as to offer.

uses neural network with three or more layers to simulate the behavior of the human brain. Its ability—allowing it to “learn” from large amounts of data. Deep learning can label, categorize, and identify patterns. It improves the ability and accuracy of AI method

Working of deep learning:

The “deep learning” refers to the depth of layers in neural network: collection of small computing units called ‘neurons. They take incoming data and learn to make decisions over time. NN are often layered deep. They makeup the backbone of deep learning algorithms. Mimic the human brain through set of algorithms.

Four main components of NN: (input, weights, and bias)

These elements work together to accurately recognize, classify, and describe objects within data. There is two main Mechanisms (forward propagation and backward propagation)

Types of neural network:

  • Perception
  • Feedforward neural networks
  • Multi-layer perceptron (MLPs)
  • Convolutional neural networks (CNN)
  • Recurrent neural networks (RNN)

The three most important types of neural networks are: Artificial Neural Networks (ANN); Convolution Neural Networks (CNN), and Recurrent Neural Networks (RNN).

Data science

The process/method of extracting ‘knowledge/insights’ from large volume of desperate data. It involves mathematics, statistical analysis, machine learning etc.

It uses machine learning (ML) algorithm and deep leaning (DL) models

Data science lifecycle: (capture, maintain, process, analyze, communicate)

Tensors:

  • they are the building blocks of “neural networks” in pytorch
  • neural networks is like “mathematical function” that takes inputs and then process it and produce output.
  • In pytorch neural networks are composed of pytorch tensors

Pytorch tensors: a ‘data structure’ that generalized of numbers and dimensional arrays in pytorch.

  • In pytorch the input(x) and the output(y) would be tensors so are the parameters.

Tensor’s operators are applied on parameters.

  • Each row of database can be created of pytorch tensor, and we input each of these rows(tensor) Into neural networks.
  • Each image can be converted to pytorch tensor and given input- mage are represented as 2D / 3D arrays.
Tags :