Artificial Intelligence, Machine Learning, Deep Learning - An Overview
Most Advanced Technologies are being used in our daily lives and we come across various terms such as Artificial Intelligence (AI), Machine Learning (ML) and Deep Learning. These Technological Advancements are used in various situations of our daily life such as while using Mail we come across Spam Filtering System, when we use OTT platforms and YouTube we come across Recommendation Systems, in our mobile phones and devices we use Google Assistant, Siri, Alexa in which it recognizes our voice and provides the results, in Autonomous Vehicles Artificial Intelligence is used, in Chatbots Natural Language Processing is used for customer service in various websites and services, Facial Recognition used in mobile phones and devices for face detection and unlocking the device and it also has various applications, etc, likewise Artificial Intelligence (AI), Machine Learning (ML) and Deep Learning are used in various situations of our daily life. A detailed explanation of Artificial Intelligence, Machine Learning and Deep Learning is discussed here with examples.
I. Artificial Intelligence (AI):
Artificial Intelligence is a field of science which is involved in building Machines and Computer Systems that can act in a way which requires Human Intelligence — which mimic Human Intelligence and perform tasks. By using Artificial Intelligence, Computer Systems mimic / simulate Human Intelligence to solve complex problems.
Artificial Intelligence is a field that encompasses many disciplines such as Data Analytics, Computer Science, Statistics, etc. For Business use in operational level, Artificial Intelligence is a set of technologies which are primarily based on Machine Learning and Deep Learning used for Natural Language Processing (NLP), Recommendation Systems, Predictions, etc.
Example - Spam Mail Filtering System:
Types of Artificial Intelligence:
Artificial Intelligence can be classified based upon Stages of development or Actions being performed (Functionalities) and what the machine can do (Capabilities).
Based on Stages of development or Actions being performed (Functionalities), Artificial Intelligence can be classified under 4 categories such as,
1. Reactive machines:
Reactive Machines are Limited AI and are the primary form of Artificial Intelligence. They do not store any past data to determine and perform future actions. Reactive Machines perform certain tasks based on preprogrammed rules, and they can’t react/perform tasks beyond those rules.
Example:
IBM’s Deep Blue was a Chess Playing Expert System which runs on IBM Supercomputer. In 1997, IBM’s Deep Blue defeated Chess Grandmaster Garry Kasparov is a Reactive Machine, which cannot refer to any of past data of playing chess or does not improve with past data, IBM’s Deep Blue can only see the Chessboard coins and reacts to it with preprogrammed rules.
2. Limited memory:
Limited Memory makes use of memory to improve its efficiency by getting trained with new data through a Deep Learning Neural Network or any other training model. Limited Memory systems performs tasks based on the past data and present data. But in Limited Memory this data isn’t saved in AI’s memory, but the AI improves over time where the AI gets trained on more data.
Example:
Self Driving Autonomous Cars.
3. Theory of mind:
Reactive Machines and Limited Memory AI Systems currently exist. But Theory of mind AI system does not currently exist and they exist as concept and Research is held for the possibilities of developing Theory of mind AI systems. Theory of mind AI systems describes as this type of AI has Decision-Making capabilities equal to Human. This type of AI can recognize, remember emotions and reacts in an environment / social situations like Humans.
4. Self-Aware:
Self-Aware AI systems are a step ahead of Theory of mind AI systems and they have the awareness of their own existence. Self-Aware AI systems have the Intellectual and Emotional Capabilities of Human. Self-Aware AI systems does not exist.
Based on what the machine can do (Capabilities), Artificial Intelligence can be classified under 3 categories such as,
1. Artificial Narrow Intelligence (ANI):
Currently Artificial Intelligence is determined as Artificial Narrow Intelligence because they can perform only specific actions based on training and pre-programming.
Example:
An Artificial Intelligence Algorithm which can perform Email Spam Filtering cannot perform Face Recognition and Natural Language Processing.
2. Artificial General Intelligence (AGI):
Artificial general intelligence (AGI) does not currently exist and this type of AI systems would have the ability to understand, learn, think, and act just like a human.
3. Artificial Super Intelligence (ASI):
Artificial Super Intelligence (ASI) also does not exist and Artificial Super Intelligence system can function in a way superior to Humans.
TYPES OF ARTIFICIAL INTELLIGENCE:
II. Machine Learning (ML):
Machine Learning is a subset of Artificial Intelligence. Machine Learning makes Computers and Machines to Learn, get Trained, improve from Data and make decisions. Instead of Pre-Programming, Machine Learning makes use of Algorithms to Analyze Data, learn from the data, get trained from the data and make decisions. Machine Leaning models / Algorithms performance increases as they are trained to more amount of data. Machine Learning Algorithm learns and get trained from large amount of data-past data to predict the future.
Types of Machine Learning Models:
1. Supervised learning:
Supervised Learning is a Machine Learning model which uses Labeled Data. In this type of Machine Learning model, it is trained with Labeled data — which means the training data is already labeled with correct output. Then the Machine Learning model analyses the data and get trained on the data. Then when a new data is given to the model, it predicts the new data based on past trained data.
Example:
When the Supervised Learning Algorithm is trained with set of Labeled BMW car images and Benz car images, then when the Algorithm is given a new car image it predicts whether the car is BMW car or Benz car.
Supervised Learning can be grouped into two types, they are,
i. Classification
ii. Regression
2. Unsupervised learning:
Unsupervised Learning is a Machine Learning model which uses Unlabeled Data. In this type of Machine Learning model, the Algorithm learns the patterns from the Unlabeled / Unstructured data without any supervision. The Algorithm learns from that Unlabeled data then it tries to find a pattern, categorize into patterns/groups based on attributes.
Example:
When the Unsupervised Learning Algorithm is given a set of car, truck and bike images, then when the Algorithm finds new patterns and groups, then when a image is given to the Algorithm it predicts whether the images is a car / truck / bike.
Unsupervised Learning can be grouped into three types, they are,
i. Clustering
ii. Association
iii. Dimensionality reduction
3. Reinforcement learning:
An Agent in Reinforcement learning learns to perform a task by by trial and error with a feedback loop until desired performance is achieved. The Agent receives reward after performing an action in the environment and then the reward measures how well the performed action is successful with respect to Goal.
TYPES OF MACHINE LEARNING MODELS:
III. Deep Learning:
Deep Learning is a subset of Machine Learning which consist of Neural Networks, these Neural Network Layers simulate the behavior of Human Brain. A Neural Network consists of Artificial Neurons / Perceptrons used to analyze data. The Input Data is first fed to the first layer of the Neural Network, then each Perceptron makes a decision and then transferring the information to further nodes of next layer. Deep Learning / Deep Neural Networks refers to the Neural Network System with more than three layers. The output of the Neural Network System accomplishes the task such as classification, etc.
Some of the most commonly used Artificial Neural Networks are,
· Recurrent neural networks (RNN)
· Long/short term memory (LSTM)
· Convolutional neural networks (CNN), etc.
Thankyou for reading this blog on Artificial Intelligence, Machine Learning, Deep Learning — An Overview!!!
Reference: https://cloud.google.com/learn/what-is-artificial-intelligence
