ARTIFICIAL INTELLIGENCE EXPLAINED IN SIMPLE WORDS
Download as DOCX, PDF1 like27 views
The document explains artificial intelligence (AI) as a digital brain mimicking biological functions, detailing three types: artificial narrow intelligence (ANI), artificial general intelligence (AGI), and artificial super intelligence (ASI). It covers machine learning (ML) and its types, including supervised, unsupervised, and reinforcement learning, emphasizing how models learn from labeled and unlabeled data through various techniques. Reinforcement learning, which involves trial-and-error and receiving feedback, has applications in diverse fields, and advancements in deep reinforcement learning have led to remarkable achievements in complex tasks.
ARTIFICIAL INTELLIGENCE EXPLAINED IN SIMPLE WORDS
- 1. ARTIFICIAL INTELLIGENCE EXPLAINED IN SIMPLE WORDS Artificial intelligence can be referred to as creating a digital brain.As we all know that humans have a biological brain which performs different body functions to survive on this planet.Now humans are trying to make a digital brain with the help of the structure of this biological brain.As we all know that our biological brain has neurons which all together perform brain functions in the same manner digital brain has also neurons(we call it neural networks) to perform different mathematical functions from simplest to complex one. Until now there are three types of Artificial Intelligence
- 2. 1.Artificial Narrow Intelligence(ANI) This branch moves around some specific topic or solve a specific problem thats why we call it Artificial Narrow Intelligence. 2.Artificial General Intelligence(AGI) Simple definition of AGI is it can do all things which we humans can do. 3.Artificial Super Intelligence(ASI) Only one type of AI exist till today that is ANI AI can be divided into ML(Machine Learning) It can be defined as ML does not need rules it can learn from given data and solve problems itself on the basis of the results of thst data. DL(Deep Learning) It is doing Machine learning with neural networks. MACHINE LEARNING There are three types of ML
- 3. 1.SUPERVISED LEARNING 2.UNSUPERVISED LEARNING 3.REINFORCEMENT LEARNING Trial & Error To explain ML we need to learn about DATA first There are two types of Data Labelled and unlabelled data(Labelled data you can refer to excel sheets and unlabelled data is without any calssification) 1.Structured Data(All data which is available in table form) 2.Unstructured Data.
- 4. There are three types of unstructured data a.Written Text b.Pictures or image(A computer does not see a image it reads numbers of image which is called matrix) c.Audio(frequencies or wavelength) SUPERVISED LEARNING Supervised learning can only be done with labelled dat.For example if i show a car picture to computer and tell him its a car and train him on 1000 pictures of different cars then computer knows that any object is car or non car.This Is Algorithm and when i test the computer by showing a random picture and get the prediction of that picture it is car or non car than it is called Model. Unsupervised Learning We use unlabelled data in unsupervised Learning.In this Learning we use clustering and dimensionality reduction with data. Reinforcement Learning In this type of Learning we use trial and error method to train the model. We use reward on right answer and penalty on wrong answer then after several steps of learning again and again the model start to give correct answers. Reinforcement learning (RL) is a type of machine learning where an agent learns to make decisions by interacting with an environment. It differs from supervised learning, where the model learns from labeled data, and unsupervised learning, which focuses on uncovering patterns in unlabeled data. In RL, the agent receives feedback in the form of rewards or penalties based on its actions, allowing it to develop a strategy or policy that maximizes cumulative rewards over time. This trial-and-error approach enables the agent to explore and exploit the environment effectively. RL has been successfully applied to a wide range of applications, including robotics, game playing, autonomous vehicles, and finance. A key challenge in reinforcement learning is balancing exploration (trying new actions to discover their effects) and exploitation (using known actions that yield high rewards). Advances in deep reinforcement learning, which combines neural networks with RL, have led to significant breakthroughs, such as training agents to play complex games like Go and Dota 2 at a superhuman level. The article will continue furthe.Keep in touch.