Artificial Intelligence for Beginners

What is Artificial Intelligence?

The answer to this question would depend on who you ask. A layman, with a fleeting understanding of technology, would link it to robots. If you ask about artificial intelligence to an AI researcher, (s)he would say that it’s a set of algorithms that can produce results without having to be explicitly instructed to do so. Both of these answers are right. So to summarize, Artificial Intelligence is:

• An intelligent entity created by humans.
• Capable of performing tasks intelligently without being explicitly instructed.
• Capable of thinking and acting rationally and humanely.

At the core of Artificial Intelligence, it is a branch of computer science that aims to create or replicate human intelligence in machines. But what makes a machine intelligent? Many AI systems are powered with the help of machine learning and deep learning algorithms. AI is constantly evolving, what was considered to be part of AI in the past may now just be looked at as a computer function. For example, a calculator may have been considered to be a part of AI in the past. Now, it is considered to be a simple function. Similarly, there are various levels of AI, let us understand those.

Why is Artificial Intelligence Important?

The goal of Artificial Intelligence is to aid human capabilities and help us make advanced decisions with far-reaching consequences. From a technical standpoint, that is the main goal of AI. When we look at the importance of AI from a more philosophical perspective, we can say that it has the potential to help humans live more meaningful lives that are devoid of hard labour. AI can also help manage the complex web of interconnected individuals, companies, states and nations to function in a manner that’s beneficial to all of humanity.

Currently, Artificial Intelligence is shared by all the different tools and techniques have been invented by us over the last thousand years – to simplify human effort, and to help us make better decisions. Artificial Intelligence is one such creation that will help us in further inventing ground-breaking tools and services that would exponentially change how we lead our lives, by hopefully removing strife, inequality and human suffering.

We are still a long way from those kinds of outcomes. But it may come around in the future. Artificial Intelligence is currently being used mostly by companies to improve their process efficiencies, automate resource-heavy tasks, and to make business predictions based on data available to us. As you see, AI is significant to us in several ways. It is creating new opportunities in the world, helping us improve our productivity, and so much more. 

History of Artificial Intelligence

The concept of intelligent beings has been around for a long time and have now found its way into many sectors such as AI in education, automotive, banking and finance, AI healthcare etc. The ancient Greeks had myths about robots as the Chinese and Egyptian engineers built automatons. However, the beginnings of modern AI has been traced back to the time where classical philosophers’ attempted to describe human thinking as a symbolic system. Between the 1940s and 50s, a handful of scientists from various fields discussed the possibility of creating an artificial brain. This led to the rise of the field of AI research – which was founded as an academic discipline in 1956 – at a conference at Dartmouth College, in Hanover, New Hampshire. The word was coined by John McCarthy, who is now considered as the father of Artificial Intelligence.

Despite a well-funded global effort over numerous decades, scientists found it extremely difficult to create intelligence in machines. Between the mid-1970s and 1990s, scientists had to deal with an acute shortage of funding for AI research. These years came to be known as the ‘AI Winters’. However, by the late 1990, American corporations once again were interested in AI. Furthermore, the Japanese government too, came up with plans to develop a fifth-generation computer for the advancement of AI. Finally, In 1997, IBM’s Deep Blue defeated the first computer to beat a world chess champion, Garry Kasparov.

As AI and its technology continued to march – largely due to improvements in computer hardware, corporations and governments too began to successfully use its methods in other narrow domains. The last 15 years, Amazon, Google, Baidu, and many others, have managed to leverage AI technology to a huge commercial advantage. AI, today, is embedded in many of the online services we use. As a result, the technology has managed to not only play a role in every sector, but also drive a large part of the stock market too. 

Today, Artificial Intelligence is divided into sub-domains namely Artificial General Intelligence, Artificial Narrow Intelligence, and Artificial Super Intelligence which we will discuss in detail in this article. We will also discuss the difference between AI and AGI.

Levels of Artificial Intelligence

Artificial Intelligence can be divided into three main levels:

• Artificial Narrow Intelligence
• Artificial General Intelligence
• Artificial Super-intelligence

Artificial Narrow Intelligence (ANI)

Also known as narrow AI or weak AI, Artificial narrow intelligence is goal-oriented and is designed to perform singular tasks. Although these machines are seen to be intelligent, they function under minimal limitations, and thus, are referred to as weak AI. It does not mimic human intelligence; it stimulates human behaviour based on certain parameters. Narrow AI makes use of NLP or natural language processing to perform tasks. This is evident in technologies such as chatbots and speech recognition systems such as Siri. Making use of deep learning allows you to personalise user experience, such as virtual assistants who store your data to make your future experience better. 

Examples of weak or narrow AI:

• Siri, Alexa, Cortana
• IBMs Watson
• Self-driving cars
• Facial recognition softwares
• Email spam filters 
• Prediction tools 

Artificial General Intelligence (AGI)

Also known as strong AI or deep AI, artificial general intelligence refers to the concept through which machines can mimic human intelligence while showcasing the ability to apply their intelligence to solve problems. Scientists have not been able to achieve this level of intelligence yet. Significant research needs to be done before this level of intelligence can be achieved. Scientists would have to find a way through which machines can become conscious through programming a set of cognitive abilities. A few properties of deep AI are-

• Recognition
• Recall 
• Hypothesis testing 
• Imagination
• Analogy
• Implication

It is difficult to predict whether strong AI will continue to advance or not in the foreseeable future, but with speech and facial recognition continuously showing advancements, there is a slight possibility that we can expect growth in this level of AI too. 

Artificial Super-intelligence (ASI)

Currently, super-intelligence is just a hypothetical concept. People assume that it may be possible to develop such an artificial intelligence in the future, but it doesn’t exist in the current world. Super-intelligence can be known as that level wherein the machine surpasses human capabilities and becomes self-aware. This concept has been the muse to several films, and science fiction novels wherein robots who are capable of developing their feelings and emotions can overrun humanity itself. It would be able to build emotions of its own, and hypothetically, be better than humans at art, sports, math, science, and more. The decision-making ability of a super-intelligence would be greater than that of a human being. The concept of artificial super-intelligence is still unknown to us, its consequences can’t be guessed, and its impact cannot be measured just yet. 

Let us now understand the difference between weak AI and strong AI. 

Weak AI	Strong AI
It is a narrow application with a limited scope.	It is a wider application with a more vast scope.
This application is good at specific tasks.	This application has an incredible human-level intelligence.
It uses supervised and unsupervised learning to process data.	It uses clustering and association to process data.
Example: Siri, Alexa.	Example: Advanced Robotics

Applications of Artificial Intelligence

Artificial intelligence has paved its way into several industries and areas today. From gaming to healthcare, the application of AI has increased immensely. Did you know that the Google Maps applications and facial recognition such as on the iPhone are all using AI technology to function? AI is all around us and is part of our daily lives more than we know it. Here are a few applications of Artificial Intelligence.

Best Applications of Artificial Intelligence in 2020

• Google’s AI-powered predictions (Google Maps)
• Ride-sharing applications (Uber, Lyft)
• AI Autopilot in Commercial Flights
• Spam filters on Emails
• Plagiarism checkers and tools
• Facial Recognition
• Search recommendations
• Voice-to-text features
• Smart personal assistants (Siri, Alexa)
• Fraud protection and prevention

Now that we know these are the areas where AI is applied. Let us understand these in a more detailed way. Google has partnered with DeepMind to improve the accuracy of traffic predictions. With the help of historical traffic data as well as the live data, they can make accurate predictions through AI technology and machine learning algorithms. An intelligent personal assistant can perform tasks based on commands given by us. It is a software agent and can perform tasks such as sending messages, performing a google search, recording a voice note, chatbots, and more. 

Goals of Artificial Intelligence

So far, you’ve seen what AI means, the different levels of AI, and its applications. But what are the goals of AI? What is the result that we aim to achieve through AI? The overall goal would be to allow machines and computers to learn and function intelligently. Some of the other goals of AI are as follows:

1. Problem-solving: Researchers developed algorithms that were able to imitate the step-by-step process that humans use while solving a puzzle. In the late 1980s and 1990s, research had reached a stage wherein methods had been developed to deal with incomplete or uncertain information. But for difficult problems, there is a need for enormous computational resources and memory power. Thus, the search for efficient problem-solving algorithms is one of the goals of artificial intelligence.

2. Knowledge representation: Machines are expected to solve problems that require extensive knowledge. Thus, knowledge representation is central to AI. Artificial intelligence represents objects, properties, events, cause and effect, and much more. 

3. Planning: One of the goals of AI should be to set intelligent goals and achieve them. Being able to make predictions about how actions will impact change, and what are the choices available. An AI agent will need to assess its environment and accordingly make predictions. This is why planning is important and can be considered as a goal of AI. 

4. Learning: One of the fundamental concepts of AI, machine learning, is the study of computer algorithms that continue to improve over time through experience. There are different types of ML. The commonly known types of are Unsupervised Machine Learning and Supervised Machine Learning. To learn more about these concepts, you can read our blog on what ML means and how it works. 

5. Social Intelligence: Affective computing is essentially the study of systems that can interpret, recognize, and process human efforts. It is a confluence of computer science, psychology, and cognitive science. Social intelligence is another goal of AI as it is important to understand these fields before building algorithms. 

Thus, the overall goal of AI is to create technologies that can incorporate the above goals and create an intelligent machine that can help us work efficiently, make decisions faster, and improve security. 

Jobs in Artificial Intelligence

The demand for AI skills has more than doubled over the last three years, according to Indeed. Job postings in the field of AI have gone up by 119%. The task of training an image-processing algorithm can be done within minutes today, while a few years ago, the task would take hours to complete. When we compare the skilled professionals in the market with the number of job openings available today, we can see a shortage of skilled professionals in the field of artificial intelligence.

Bayesian Networking, Neural nets, computer science (including knowledge about programming languages), physics, robotics, calculus and statistical concepts are a few skills that one must know before deep diving into a career in AI. If you are someone who is looking to build a career in AI, you should be aware of the various job roles available. Let us take a closer look at the different job roles in the world of AI and what skills one must possess for each job role. 

1. Machine Learning Engineer

If you are someone who hails from a background in Data Science or applied research, the role of a Machine Learning Engineer is suitable for you. You must demonstrate an understanding of multiple programming languages such as Python, Java. Having an understanding of predictive models and being able to leverage Natural Language Processing while working with enormous datasets will prove to be beneficial. Being familiar with software development IDE tools such as IntelliJ and Eclipse will help you further advance your career as a machine learning engineer. You will mainly be responsible for building and managing several machine learning projects among other responsibilities.

As an ML engineer, you will receive an annual median salary of $114,856. Companies look for skilled professionals who have a masters degree in the related field and have in-depth knowledge regarding machine learning concepts, Java, Python, and Scala. The requirements will vary depending on the hiring company, but analytical skills and cloud applications are seen as a plus point. 

2. Data Scientist 

As a Data Scientist, your tasks include collecting, analyzing, and interpreting large & complex datasets by leveraging machine learning and predictive analytics tools. Data Scientists are also responsible for developing algorithms that enable collecting and cleaning data for further analysis and interpretation. The annual median salary of a Data Scientist is $120,931, and the skills required are as follows: 

• Hive
• Hadoop
• MapReduce
• Pig
• Spark
• Python
• Scala
• SQL 

The skills required may vary from company to company, and depending on your experience level. Most hiring companies look for a masters degree or a doctoral degree in the field of data science or computer science. If you’re a Data Scientist who wants to become an AI developer, an advanced computer science degree proves to be beneficial. You must have the ability to understand unstructured data, and have strong analytical and communication skills. These skills are essential as you will work on communicating findings with business leaders. 

3. Business Intelligence Developer 

When you’re looking at the different job roles in AI, it also includes the position of Business Intelligence (BI) developer. The objective of this role is to analyze complex datasets that help us identify business and market trends. A BI developer earns an annual median salary of $92,278. A BI developer is responsible for designing, modelling, and maintaining complex data in cloud-based data platforms. If you are interested to work as a BI developer, you must have strong technical as well as analytical skills.

Having great communication skills is important because you will work on communicating solutions to colleagues who don’t possess technical knowledge. You should also display problem-solving skills. A BI developer is typically required to have a bachelor’s degree in any related field, and work experience will give you additional points too. Certifications are highly desired and are looked at as an additional quality. The skills required for a BI developer would be data mining, SQL queries, SQL server reporting services, BI technologies, and data warehouse design. 

4. Research Scientist 

A research scientist is one of the leading careers in Artificial Intelligence. You should be an expert in multiple disciplines, such as mathematics, deep learning, machine learning, and computational statistics. Candidates must have adequate knowledge concerning computer perception, graphical models, reinforcement learning, and NLP. Similar to Data Scientists, research scientists are expected to have a master’s or doctoral degree in computer science. The annual median salary is said to be $99,809. Most companies are on the lookout for someone who has an in-depth understanding of parallel computing, distributed computing, benchmarking and machine learning. 

5. Big Data Engineer/Architect 

Big Data Engineer/Architects have the best-paying job among all the roles that come under Artificial Intelligence. The annual median salary of a Big Data Engineer/Architect is $151,307. They play a vital role in the development of an ecosystem that enables business systems to communicate with each other and collate data. Compared to Data Scientists, Big data Architects receive tasks related to planning, designing, and developing an efficient big data environment on platforms such as Spark and Hadoop. Companies typically look to hire individuals who demonstrate experience in C++, Java, Python, and Scala. 

Data mining, data visualization, and data migration skills are an added benefit. Another bonus would be a PhD in mathematics or any related computer science field.

Advantages of Artificial Intelligence

Just as it is the case with most things in the world, AI has its pros and cons. First, let us understand the advantages of artificial intelligence and how it has made our lives easier compared to the earlier times. 

• Reduction in human error
• Available 24×7
• Helps in repetitive work
• Digital assistance 
• Faster decisions
• Rational Decision Maker
• Medical applications
• Improves Security
• Efficient Communication

Let’s take a closer look at each of the aforementioned points. 

1. Reduction in human error

All decisions taken in an AI model are taken from previously gathered information after having applied a set of algorithms. This enables the errors to be reduced, and the chances of accuracy increase with a greater degree of accuracy. In the case of humans performing any task, there’s always a slight chance of error. Since we are capable of making errors, it is better to make use of programs and algorithms through AI as they lower the chance of errors. 

2. Available 24×7

Artificial intelligence models are built to work 24/7 without any breaks or boredom. When compared to an average human who can work for six to eight hours in a day, this is significantly more efficient. Human beings do not have the capacity to work for longer durations as we would require rest and time to rejuvenate. Thus, AI is available 24/7 and improves efficiency to a greater extent. 

3. Helps in repetitive work

Artificial Intelligence can productively automate mundane human tasks. It can help us in becoming increasingly creative – right from sending a thank you mail to decluttering or answering queries. It can also help us in verifying documents. A repetitive task such as making food in a restaurant or a factory can be ruined because humans become tired or uninterested after a long duration of work. AI can help us in performing these repetitive tasks efficiently and without error. 

4. Digital assistance

Several organisations who are highly advanced make use of digital assistants to interact with users. Doing so helps the organisation to save costs on human resources. Digital assistants such as Chatbots are typically used in an organisations website to answer user queries. It also provides a smooth functioning interface and good user experience. Chatbots are a great example of the same. Read here to know more about how to build an AI Chatbot.

5. Faster decisions 

AI, alongside other such technologies, can help machines take faster decisions when compared to an average human being. This helps in carrying out actions quickly. This is because, while making a decision, humans tend to analyse factors through emotions as opposed to AI-powered machines that deliver programmed results quickly.

6. Rational Decision Maker

We as humans may have evolved to a great extent technologically, but when it comes to decision making, we still allow our emotions to take over. In certain situations, it is really important to take quick, efficient and logical decisions without our emotions coming into the picture. AI-powered decision making is controlled by AI algorithms, and thus, there is no scope for any emotional discrepancy. Rational decisions with the help of AI ensures that efficiency will not be affected, and also increases an organisation’s productivity level. 

7. Medical applications

Among all other advantages of AI, one of the greatest applications in its use in the medical field. Doctors can assess their patients’ health risks with the help of AI-powered medical applications. Radiosurgery is being used to operate on tumors in such a way that it won’t damage surrounding tissues and cause any additional issues. Medical professionals have been trained to use AI for surgery. They can also help in efficiently detecting and monitoring various neurological disorders and stimulate the brain functions. 

8. Improves Security

As technology continues to advance, there is a higher chance of people using it for unethical reasons such as fraud or identity theft. If used in the right manner and for the right reasons, AI can prove to be a great resource in improving our organisation’s security. AI can be used to protect our data and finances. AI is being implemented majorly in the field of cybersecurity. It has transformed our ability to secure our personal data against any cyber-threats or attacks of any form. Read further to know about AI in Cybersecurity and how it helps, here.

9. Efficient Communication 

People from different parts of the world speak different languages and thus, find it hard to communicate with each other. When we look at the past, we see how human translators would help people communicate with each other if the other person did not understand the same language as us. Such problems do not occur if we make use of AI. Natural Language Processing allows systems to translate words from one natural language to another, thus eliminating the middleman. One of the best examples of this is Google translate, and how it has advanced over time. Now, it provides audio examples of how words/sentences should be pronounced. Thus, improving our accuracy and ability to communicate effectively.

Disadvantages of Artificial Intelligence

Now that we have understood the advantages of AI, let us take a look at a few disadvantages. 

• Cost overruns
• Dearth of talent
• Lack of practical products
• Lack of standards in software development
• Potential for misuse
• Highly dependent on machines
• Requires Supervision

Let’s take a closer look at the disadvantages of AI. 

1. Cost overruns

The scale of operations of an AI-powered model when compared to software development is massively higher. Due to this, the resources required increase at a much higher rate. This pushes the cost of operations to a higher level.

2. Dearth of talent 

AI is still a field which is developing. Thus, finding professionals who are equipped with all the required skills is not easy. There is a gap between the number of jobs available in the field of AI vs the skilled workforce in the field. Hiring someone who possesses all the necessary skills further increases the costs incurred by an organisation.

3. Lack of standards in software development

The true value of Artificial Intelligence lays in collaboration when different AI systems come together to form a bigger, more valuable application. But a lack of standards in AI software development means that it’s difficult for different systems to ‘talk’ to each other. Artificial Intelligence software development itself is slow and expensive because of this, which further acts as an impediment to AI development.

4. Potential for Misuse

AI has the potential to achieve great things, and has massive power in the market today. Unfortunately, with great power comes the potential of misuse. If the power of AI falls in the hands of a person who has unethical motives, there is a higher chance of misuse.

5. Highly dependent on machines

Applications such as Siri and Alexa have become part of our everyday lives. We are highly dependent on these applications and receive assistance from these applications, thus reducing our creative ability. We are becoming highly dependent on machines and losing our on learning simple skills, thus becoming lazier. 

6. Requires Supervision

Making use of AI algorithms has a lot of advantages and is highly efficient. But it also requires constant assistance and supervision. These algorithms cannot work without us programming them and checking if they are functioning in the right manner or not. One example is Microsoft’s AI chat-bot named ‘Tay’. Tay was modelled to speak like a teenage girl by learning through online conversations. But since it was programmed to learn basic conversational skills and didn’t know the difference between right and wrong, it went ahead and tweeted highly political and incorrect information because of internet trolls.

Future of Artificial Intelligence

We have always been fascinated by technological changes. Currently, we are living amidst the greatest AI advancements in our history. Artificial Intelligence has emerged to be the net greatest advancement in the field of technology. This has not only impacted the future of every industry, but has also acted as the driver of emerging technologies such as big data, robotics and IoT. At that rate at which AI is advancing, there is no doubt that it will continue to flourish in the future. Thus, we can say that AI is a great field to enter as of 2020. With the advancement of AI and its technologies, there will be a greater need for skilled professionals in this area.

An AI certification will give you an edge over other participants in the industry. As Facial Recognition, AI in Healthcare, Chat-bots continue to show growth, now would be the right time to work towards building a successful AI career. Virtual assistants are already part of our everyday life without us knowing it. Self-driving cars by Tech giants like Tesla have shown us a glimpse of what the future will look like. There are so many more advancements to be discovered, this is only the beginning. According to the World Economic Forum, 133 million new Artificial Intelligence jobs are said to be created by Artificial Intelligence by the year 2022. The future of AI is definitely bright.

A simple Artificial Intelligence mini-project

Before moving on to the project, I would suggest going through this Machine learning Tutorial if you are not familiar with Machine learning at all. It would also help you with this project if you know about the Logistic Regression algorithm.

Zoo Animal Classification

In this mini-project, we will use different algorithms that come under the Machine learning domain of Artificial Intelligence to classify animals in a zoo, based on their attributes. We are going to use this dataset from Kaggle which consists of 101 animals from a zoo. There are 16 variables with various traits to describe the animals. The 7 Class Types are: Mammal, Bird, Reptile, Fish, Amphibian, Bug and Invertebrate.

The purpose of this dataset is to be able to predict the classification of the animals based on the variables. You can also find the information about the various attributes used in this dataset from the download page linked here.

Output:

Output: 

training accuracy : 1.0

testing accuracy : 0.9215686274509803

As you can see, the model performed exceptionally well by getting 92% accuracy on the testing data. Now, if you are given the attributes of any of the animal in the above dataset, you can classify it with the help of the above model.