AI in Action

AI in Action

We humans in today’s world have been hearing the word called “Artificial Intelligence”. But most of us don’t know what does Artificial Intelligence exactly means, and moreover, why it is unavoidable in today’s real-world of technology and how helpful is AI in industrial science systemization.

What is Artificial Intelligence?

The capability of a mineral machine to mimic the intelligent human behavior. Artificial Intelligence has previously predicted the upcoming future and makes the globe to adopt technological changes.

In simpler terms, It is “machine intelligence”, building or designing artificial systems that think exactly or more effectively than that of what humans do.

Why is AI needed & Why is AI essential?

AI replaces “repetitive jobs”. It literally means that these kinds of monotonous jobs will be automated, like what robots and makes the task simpler.AI creates an abundance of new jobs than ever before. According to this outlook, Artificial Intelligence will be the most noteworthy employment engine that the globe has ever seen. AI abolishes repetitive jackleg jobs and effectively creates massive high-skilled job opportunities that will span all sectors.

The work that technology needs to do is increasing day-to-day, the need for Artificial Intelligence also increases. So it is an ideal objective to automate the colorless routine work. This saves the manpower of the organization and also increases productivity on a larger scale. In addition, through this machine intelligence, the company gets skilled and accomplished persons for the development of their organization. The companies today want to mechanize the regular routine work and they automate those works through simple programming.

When did AI evolve?

Deep Blue is the first artificial chess machine that defeated the World Chess Champion Garry Kasparov in a six-game match in 1997. The “DEEP BLUE” was first built by IBM scientists. It consists of a single-chip chess search engine.

S2: FATHER OF ARTIFICIAL INTELLIGENCE: John McCarthy

How do AI works? 

The goal of AI is to build systems that can function intelligently. Computers fundamentally need to be told what should they do. So how can we get it to make its own informed decisions? This is an issue of machine learning.AI is implemented in so many different ways. The solution for this starts with an algorithm.

There are different machine learning algorithms and protocols that are applied, but to get the best idea of how this works we get into the topic of biologically inspired computing. Building something to think is to stimulate some kind of information processing machine that they can make different decisions on the basis of neural networks. Neural networks are built up from systems called perceptual inner a kind of artificial neurons. Each neuron takes in several binary inputs x1, x2, x3 so on and produces an output Y. Artificial Intelligence works mainly on the following three broad techniques.

• Symbolic AI
• Data-Driven
• Future development

Symbolic AI is the most classical section of Artificial Intelligence. Symbolic AI possesses an outstanding interpretation potentiality. In earlier times, machines were used to check only the problem-solving capabilities with logic and reason. Data-driven AI customizes consulting firms with advanced analytics. And this is applied in NLP (Natural Language Programming). Artificial Intelligence uses historical data and algorithms like Naive Bayes, Decision Tree, Random Forest, Logistic Regression, Support Vector Machines, K-nearest neighbors to build a model called “PROPENSITY MODEL”.

The propensity model makes predictions like spur rotation, human augmentation, hyper-automation, multi-experience, and so on.

Types of AI:

Based on ability:                              Based on functionality:

1.Narrow AI                                            1.Reactive Machines AI

2.General AI                                           2.Limited Memory AI

3.Super AI                                              3.Theory of mind AI and Self-awareness

Narrow Artificial Intelligence(NAI):

In “NAI or WEAK AI”, the machines do not possess their own conscience and the lack of thinking ability. They simply perform the task given to them outlined with predefined functions.

Eg: Siri, Alexa, Alpho, Sophia, self-driving cars then on. All the AI systems engineered until the date comes beneath NAI.

General Artificial Intelligence(GAI):

“GAI or STRONG AI” is that the stage of AI, where machines possess the flexibility to assume and create choices rather like humans. We’ll presently be able to produce a machine that is as good as humans. Strong AI takes into account as a threat to human existence by scientists.

Super Artificial Intelligence(SAI):

Super AI is that the stage of AI wherever the potential of machines can exceed human intelligence. During this stage of AI, machines are mentally keen or faster than humans. This kind of AI might end in the extinction of humans.

Reactive machines AI:

In Reactive machines, AI, the system takes the present data of any current situation into the account. In this type, the machines fail to predict future actions.

Limited memory AI:

In the case of Limited memory AI, the machines store both the present and past data in its memory. And this kind of memory storage makes the inbuilt system to predict future actions accurately.

Theory of mind AI:

Theory of mind AI deals with emotional intelligence so that human’s beliefs, feelings, and thoughts can be better comprehended by the machines.

Self-aware AI:

The self-aware AI machines that have their own consciousness and become aware of thyself. But, this kind of AI does not exist to date.

Machine learning:

Machine learning is the study of algorithms that get upgraded through test and trial case. Machine learning algorithms create the calculation and analytical model based on the given sample data using “training data”(70%) and “test data”(30%).

Deep learning:

Deep Learning comes under Machine Learning; Deep learning is a subset of ML where machines use artificial neural networks. Let us try to figure this out with a few examples:

Sample 1:

Let us try and understand how you’ll get a square from other shapes:

    =>  Sides=4, Closed, Perpendicular, Equal sides 

The first thing is that we have to check whether there are four lines associated with the given figure. If yes, we start moving to the second step that the lines are connected with each other or not and forms a closed figure. If yes, we check whether the lines are perpendicular to each other and we finally check all the four sides are equal in length.

• We took a complex task of identifying the square from the given figures.
• We broke the complex one into simpler tasks
• Deep Learning also does the same task but in a wide range larger scale.

Sample 2:

Let us take an example of a machine that recognizes the animals.

The task of the machine: To identify precisely whether the given image s3 is a dog or cat.

We’ll first instruct the machine with features such as whether the given animal has whiskers or not, next to the machine checks if the animal possesses pointed ears. Then finally check whether the tail is curved or straight.

In simple words, We’ll define the facial features and let the machine to classify the animal based on the features and train the model to predict the animal.

s3: Dog vs Cat

What does deep learning do?

It takes this model to move a step forward. Deep learning finds out the required features for further classification and verification of the system. In Machine Learning where we had to humanly give the features and instructions to predict the model.

Applications of AI:

Agriculture:

• AI shows improvement in gain crop, yielding and predicts the time taken for a crop to be ripe.
• Increase the efficiency of farming.

Hospitals and medicines:

• Analysis of heartbeat
• Providing consultations
• Drug creation
• Using avatars instead of patients for clinical treatment
• Automatic calculation of bone age using an X-ray of a hand.
• Virtual nursing assistant
• Robotic surgery

S4: Applications fo AI

• Cybersecurity Intelligence, Inhuman resource management, AI in e-commerce and finance, Gaming, Chatbot, Space Exploration, Social media, and entertainment.

Sources:

S1: https://norrismclaughlin.com/blb/wp-content/uploads/sites/5/2019/06/artificial-intelligence-hologram.jpg

S2: https://media.wired.com/photos/59327c7752d99d6b984dee6e/master/w_542,c_limit/john-mccarthy.png

S3: https://d1jnx9ba8s6j9r.cloudfront.net/blog/wp-content/uploads/2018/06/cat-vs-dog.png

S4: https://qph.fs.quoracdn.net/main-qimg-759e9ba390fe63b73ccfa514cbff69f9