Types of Artificial Intelligence You Must Know

Download ebook

Ever watched the 2004 box office hit, ‘I, Robot’? Brilliant movie with a brilliant performance by actor Will Smith. Set in 2035, Will Smith plays a cop who is averse to artificial intelligence, as he investigates the murder of a scientist with the help of a prototype robot. The movie depicts highly intelligent robots managing day to day services and operations.

This futuristic movie predicts how our life may be hacked by artificial intelligence in a non-invasive way. There are hundreds of movies portraying the efficiency of machines in managing activities of daily living.

I infer, our future too will be controlled by machines soon. Today, most of the organisations have incorporated artificial Intelligence in their operations to improve and gain ground.

Like robotics, there are other forms of AI that we are already using, such as automation, self-driving cars, machine learning, and chatbots. As kids, we have wished for such a technology, yearning to experience them in our lifetime. Perhaps we have reached the verge.

Here, we will explore the various kinds of Artificial Intelligence and much more, but first let’s answer the most important question :

Are you an AI and Machine Learning enthusiast? If yes, the AI and Machine Learning course is a perfect fit for your career growth.

What is Artificial Intelligence?

Artificial intelligence refers to the branch of computer science, which helps build intelligent machines to execute tasks that need human intelligence.

In short , these smart machines simulate human intelligence.

Some applications of AI are Siri, Alexa, expert systems, NLP, machine vision and speech recognition. These applications continually evolve to make life better. 

There are mainly 4 types of artificial intelligence:

1. Reactive machines
2. Limited memory
3. Theory of mind
4. Self-awareness

1. Reactive machines

Reactive machines only implement the fundamental principles of AI. They use their intelligence to detect and react to the existing situation. In real time, they cannot create memories and use them in decision-making.

In other words, they function using the stimulus and response theory.

This means they can perform only limited tasks, however this doesn’t provide any cost-cutting benefit. Instead, every time they encounter the same stimulus, they repeat their response. This behaviour leads to reliability, as their reactions don’t vary with time .

Deep Blue is a reactive machine designed by IBM in the 1990s as a chess- playing supercomputer. This reactive machine has even defeated grandmaster Gary Kasporov.

However, this AI system only knew the rules of chess and recognised the best placement possibilities of each piece. It made moves using logic rather than determining the potential move of the opponent beforehand. Owing to its nature, this supercomputer viewed every move as a separate not connected to the previous move.  

Though a reactive machine has its limitations, it performs some complexities. As it cannot be altered, it provides reliability while performing repeatable tasks.

2. Limited Memory

This kind of artificial intelligence is more complex than the reactive machine. Since it has a memory, it can store data and make use of its experience while deciding potential moves. It looks for clues in the past to predict its future decisions.

A team trains this AI model repeatedly to analyse and use the data in a given environment. A good example is self-driving cars.

When using this AI model, six steps repeat as a cycle :

• Creation of training data
• Creation of machine learning model
• This model must make predictions
• It should receive human and environmental feedback
• The feedback must be stored as data by the model

These models use the limited memory artificial intelligence:

• Reinforcement memory uses the trial and error method as it learns to make predictions.
• Long short-term memory, when predicting a sequence, uses the most recent memory and uses the data from the past to shape conclusions.
• Evolutionary Generative Adversarial Networks (EGAN) utilises the evolutionary mutation model, where its decision is based on its experience. It constantly looks for a better solution by using simulations and statistics.

3. Theory of mind

This is just a futuristic theoretical concept. We haven’t achieved this yet. This type of artificial intelligence can identify human emotions by using social intelligence.

We function using our thoughts, feelings, and memories, which influence our behaviour. And the theory of mind is based on this concept, hoping that machines could also replicate human emotions.

Thus, by comprehending emotions and human intentions, these machines can predict behaviour. By processing these psychological concepts,a relationship between humans and artificial intelligence can be formed.

These highly intelligent supercomputers base their decisions by understanding other’s thoughts and mind.

An example of this technology is Sophia, the famous robot created by Hanson Robotics, who is often interviewed by the press. They try to display the capacity of an ever-evolving robot.

However, Sophia is far from the theory of mind psychology, as she can only determine emotions and hold conversations. She also possesses image recognition and can well interact with humans while practising appropriate facial reactions .

4. Self-awareness

This AI machine is termed as the most complex supercomputer and perhaps the supreme goal of developers.

Under this theory, AI machines have self-awareness, which helps them understand their own existence, the existence of other beings, and the emotions of others.

They possess human-level consciousness and can interpret its internal state and the feelings of others around them.

For example, these machines can conclude that someone is angry when they yell and also detect their need when they want something.

This state can be established once the theory of mind conceptualises. This requires the human level of consciousness. First, researchers have to understand consciousness and then replicate it in machines.

At present, this artificial intelligence model is science fiction and could even remain so!

Our current focus is on the development of machine learning. Machines with a memory, which can learn from previous experiences and then apply it when making decisions, will be the foundation of the evolution of AI.

This development can lead to AI innovation , which can entirely change the world.

Let’s continue to understand artificial intelligence better.

How does AI work?

Artificial Intelligence requires a specialised hardware and software for its algorithms. It uses over one language like Python, R and Java and more.

When large volumes of data are fed into the systems, then these machines analyse the data to recognise patterns and correlations which they utilise to make predictions.

For example, a chatbot, when fed with texts, exchanges information with people or an image recognition application, recognises images from among millions and describes it.

There are 3 skills that AI programming requires:

Learning processes

AI programming focuses on data processing and thus requires instructions.These instructions take the form of algorithms for step by step rules on how to complete a specific task.

Reasoning processes

AI programming also focuses on selecting a specific algorithm for the right output.

Self correction processes

AI programming can also make improvements in the algorithms from now and again to ensure best results. 

What are the examples of AI technology used today?

Here are six examples where AI technology is incorporated:

Automation

It is true, artificial intelligence can get more work done. RPA (Robotic Process Automation) gets repetitive and data processing jobs done automatically.

These jobs are traditionally done by humans, but by combining it with AI tools and machine learning , bigger volumes of enterprise jobs are automated.

Machine learning

This is getting machines to complete tasks with feeding data. Here, computers use predictive analysis.

Machine vision

Signature identification and medical image analysis is an example of machine vision. This technology enables a machine to see by capturing and analysing images, digital signal processing and analog to digital conversion.

Though it is compared to human vision, it can be programmed to see through walls as it is not bound by biology.

Natural Language Processing

The processing of human language by machines is NLP.

Spam detection and segregating it as junk is an example of this. Even speech translation and voice recognition are classic models of NLP.

Robotics

This focuses on the creation of robots. Robots can undertake tasks that are difficult to achieve, like carrying heavy objects. Even jobs that require consistency, like assembly lines, are performed by robots.

Self driving cars

Self driving cars operate by using the right combination of computer vision, image recognition and deep learning. These auto-pilot vehicles are programmed to stay in the same lane and avoid all obstacles.

What are the applications of AI?

Artificial Intelligence is used in various fields:

Healthcare

Machine learning is leaning towards faster diagnosis , much faster than humans. IBM Watson is the best example. It responds to questions as it can understand natural language because it extracts data and creates perfect schema based on a hypothesis.

There are also virtual health assistants and chatbots to help patients with all the medical information, billing, scheduling appointments and much more.

Presently, AI technologies are implemented to prevent and fight the COVID-19 pandemic.

Business

In order to serve customers better, machine learning algorithms are integrated into CRM( Customer Relationship Management) platforms.

Chatbots and automated job positions are incorporated into the websites for service assistance.

Education

AI assists educators by assessing and grading students and thereby saving time. It even tracks the learning patterns of students and provides additional support.

Finance

AI applications give financial advice by collecting personal data, help in buying homes and also trade on stock markets .

It is an enormous loss for many financial institutions.

Law

Screening through documents is a tedious process for humans. AI comes to the rescue by automating documentation and helps save time and improve service.

Law firms have begun using machine learning, NLP, and machine vision for various tasks, such as predicting outcomes, interpret requests and get information from documents, respectively.

Manufacturing

The manufacturing industry is the forerunner for introducing robots into the work system. Early on, there were industrial robots focusing on single tasks , separated from human workers, but today we have cobots.

These cobots co-work with their human counterparts in factories and warehouses by performing most of the work. However, these cobots are much smaller these days and they are specialised in multi-tasking too.

Banking

Similar to most businesses, banks too have employed chatbots to provide appropriate information to customers and perform banking transactions that do not require human intervention.

Virtual assistants are making way for cost cutting techniques. Overall, AI is helping banks in setting credit limits and also in giving credit.

Transportation

Not only do AI machines autopilot vehicles, they also make ocean shipping better, manage traffic, manage flight delays and more.

Security

Organizations use AI machines to keep anomalies and security threats at bay. Machine learning detects suspicious activities by analysing data and using logic, then alerts us if there are any cyber attacks. Artificial intelligence warns us much before humans can identify any malicious attacks on the system.

What are the advantages and disadvantages of artificial intelligence?

Well, after studying the uses and applications of AI in our systems, we derive AI helps in faster processing of data.

When there is an enormous volume of data to process ,it could burn out humans. Machine learning helps process information quickly while even making accurate predictions.

The advantages are:

• Virtual agents powered by AI are 24*7 available.
• Large amounts of data are processed easily and faster.
• Results are reliable and consistent.
• Meticulous at all jobs.

The disadvantages of AI are:

• Requires special expertise.
• Expensive
• AI machines know only what they’ve been programmed to perform.
• Dearth of AI workers to build tools for machine learning.

Will AI cause concerns for the future?

The answer to this question is quite relative. There are generally 3 kinds of people in the world, first, people who support AI and are convinced about the perks, second, people who are hostile towards the idea and think machines are unfavourable. Last, a group of people who live in complete blissful ignorance, come what may .

Anyone who has watched enough sci-fi movies must know that all hell breaks loose when machines gain consciousness. The takeover is just not physical, but mental and emotional.

So, is it really that bad?

There is no single answer, but it is better to make a smart move by carefully examining the creation and utilisation of a self-aware machine.

Our foresight warns us about the doomsday theory where machines displace humans and rob them off their jobs and income. Also, the idea of subjugating a self-aware and fully conscious being, seems unethical.

There is also the concern of creationists and certain religious groups who do not welcome this theory . It is believed ,because of further advancement of technology and the widespread use of AI, humans would have to adapt to challenges where they submit to machines and laws created by them . Humans could be superseded.

Nevertheless, people who favour AI support it because of the countless advantages.

Let’s study how is AI used:

There are two categories of artificial intelligence:

Narrow AI

Also referred to as the weak AI, it works within a range and it simulates human intelligence. It is designed to complete a specific task extremely well;

We use narrow AI all around us each day and are the most successful AI. Examples of narrow AI are:

• Alexa; Siri
• Google assistant
• Image recognition software
• Self-driving cars
• IBM’s Watson

Machine Learning and Deep learning

A majority of narrow AI is powered by machine learning and deep learning.

Machine learning learns how to improve at a task progressively without being programmed for it. It uses computer data and statistics to advance at that task and eliminates the need for written codes for it. It could be both supervised or unsupervised learning .

Deep learning is machine learning with underlying neural network architecture.This contains numerous layers which process data and allow ‘deep’ learning.

Artificial General Intelligence

Researchers are still on a quest for AGI, as it requires machines with human intelligence to perform tasks. Creation of a machine with cognitive abilities seems profound as it requires a universal algorithm to perform in any environment.

If you AI interests you and you want to pursue it please log on to :

https://www.henryharvin.com/analytics-academy

Meanwhile here are some insights into the course:

AI at Henry Harvin

Henry Harvin Education ,the prestigious institution, has its own analytics academy where you can pursue AI and be a certified artificial intelligence practitioner.

This course is one of the best in the country, with a well-planned curriculum. This curriculum will help you learn the appropriate techniques from the experts. And it will make you eligible for jobs from across the world.

HH is well known for its practical training and here too they provide firsthand CAIP courses with live-projects with popular GCAO pedagogy.

They provide total job support and also brush up on your skills for the interview.

Post training, you also get 12 projects for a year so you can showcase your experience.

The trainers here are field experts with enormous teaching experience.

The duration of the programme is 32 hours + 50 hours of e-learning. You can also become an intern at Henry Harvin or their partner firms.

Not only this, but they also provide a 100% money-back guarantee where you will be refunded your entire fee if you are not satisfied with the research and writing course after the first session.

What would you be learning at the analytics academy?

• The purpose , meaning and applications of artificial intelligence.
• Fundamental concepts of Machine learning and deep learning.
• Clustering and classifying algorithms in business applications.
• Supervised, unsupervised, and revised learning.
• Machine language and AI theories.

Henry Harvin has various courses on artificial intelligence . Kindly log on to :

https://www.henryharvin.com/analytics-academy

Conclusion

As we see today, AI machines have become a necessity. With the abundance of data and knowledge, manual work has become tedious. The intellectual demands are also on the rise, in order to satisfy the various demands ,we need to rely on machines. However, to what extent do we rely on machines is the question? Where do we halt?

Answering these questions isn’t simple because there are millions of people with different agendas up their sleeves and unifying their thought patterns is impossible. This brings us to the ethical issues in artificial intelligence.

But first, why is AI important?

Because AI has given a newer dimension to enterprise operations, and performs tasks better than humans , it has become indispensable. Specifically, jobs which are detail-oriented and repetitive as AI machines perform them quickly with fewer errors.

This level of efficiency has therefore paved the way for many new businesses to pop up. For example, Uber has used machine learning algorithms successfully in connecting riders to their taxis. Also google uses AI extensively to serve people better online. Many organisations have invested in AI to gain advantage over the competitors.

Fraud detection, logistic optimisation, art composition, translations, researching: these machines have transformed our lives. Microsoft, IBM, Google, Amazon swear by these machines. Despite these advantages, there are risks involved which humans cannot compromise on.

These are some factors that we need to consider:

1. Automation of jobs has given rise to unemployment issues.
2. By reducing the human workforce, there is an unequal distribution of wealth.
3. AI bots are becoming more intelligent and are improving conversation styles each day. Knowingly or unknowingly, this is affecting humanity as they are creating relationships . The behaviour patterns are influenced as the interaction between machines and humans is increasing.
4. Machine learning comes when humans train these machines, but it isn’t possible to teach everything under the sun at once. Hence, creating loopholes where machines can still make mistakes.
5. Are we indeed safe from autonomous weapons? Robots are replacing soldiers. Is that really making the world safer? Can we protect our planet from unintended consequences?
6. Defining humane treatment of robots and robot rights. Do they make sense?

Some of these questions are about reducing risks while some are to ease human suffering.

Through Artificial intelligence we can better our technology and make life comfortable. Somehow, its responsible implementation is on all of us and we must mindfully exploit this futuristic system.

FAQs

Download ebook