Towards AI — An Introduction

The voyage begins …

First and foremost, I want to appreciate people near to me for encouraging me to publishing this educational article. By being aware of my social responsibility and giving back to the community, I wish to pass on the torch of knowledge. If this article sparks curiosity, please give it a 👏 or share it with other social media.

Image credit — Colgate.Edu

In the United States, the Thomas Jefferson Building of the Library of Congress is crowned with the Torch of Knowledge.

I am presenting this article as an introductory part of the series intended for the Artificial Intelligence (AI) technology enthusiasts with a computer science background who are new to the domain and would like to get started understanding terminologies used and gain deeper understanding of subtopics down the line. This article serves as an educational material at the end of which, one will be able to clearly identify various types of AI, its subfields, its components along with threats and benefits posed by AI to the people, including its public and private domain usage. As a subject matter expert, I strongly believe that AI can benefit us more than we thought possible and it’s important to learn AI as a technology to survive in this cutting edge technological world. So let’s begin our journey.

Image Credit — ScienceDirect

“A computer would deserve to be called intelligent if it could deceive a human into believing that it was human.” — Alan Turing

This is an article to outline and understand what lies underneath the term AI (Artificial Intelligence) in simplistic terms.

AI is a field with a combination of computer science, mathematics, and data engineering to help machines mimic human capabilities.

Table of Contents

1. Types of AI
2. Subfields of AI
3. Components of AI
4. AI and Society
5. Can AI help the Government?
6. Threats posed by AI
7. Key Takeaways
8. What’s Next?
9. Further Reading

1. Types of AI

1.1 Artificial Narrow Intelligence — (ANI)

This is also referred to as weak AI because it showcases intelligence in a given domain such as chess playing, sales prediction, movie suggestions, weather forecasts etc. These are part of reactive machines and to a certain extent belong to limited memory machines.

1.2 Artificial General Intelligence — (AGI)

This is also referred to as strong AI because it operates at the same level as humans do such as solve complex problems, control itself autonomously with thoughts, worries, feelings, emotions and so on but this is yet elusive to achieve completely by a machine in similarity to humans. These are part of limited memory machines and expand into theory of mind systems.

1.3 Artificial Super Intelligence — (ASI)

This is revolving around the concept that it exceeds human capabilities across several domains. Again hypothetical to be achieved yet. These are also a part of Theory of mind and Self aware systems.

2. Subfields of AI

The subfields of AI include Machine Learning, Knowledge Based Systems, Computer Vision, Robotics, Natural Language Processing, Automated Planning and Scheduling along with Optimization.

2.1 Machine Learning

It is concerned with the design and use of computer systems to learn from past and current data for modeling, control and prediction using statistical methods. They include subcategories 1. Supervised Learning and 2. Unsupervised Learning.

2.2 Knowledge Based Systems

It is a branch of AI that deals with decision making based on existing domain knowledge. Example of it includes Expert systems which are task — specific and contains domain knowledge pertaining to the task at hand.

2.3 Computer Vision

It is concerned with artificial simulation of the human visual system by being multidisciplinary.

2.3 Robotics

This discipline deals with automated devices that interact with the real world via physical activities. Concept is to mimic human physical capabilities.

2.4 Natural Language Processing

They deal with the idea of mimicking human linguistic capabilities. Examples include speech-to-text and vice versa.

2.5 Optimization

This deals with optimizing the choice of decision or outcome given constraints. It is a mathematical discipline and is a lifelong phenomena.

3. Components of AI

Major components include Learning, Reasoning, Planning, Action, Perception, Knowledge representation and Communication where the idea is to mimic human senses.

4. AI and Society

AI is already proving its worth via applications such as predictive prescriptions, affordable housing, voice assisted smartphone, handwriting recognition mail delivery, financial trading, smart logistics, spam filtering, language translation and so on to connect the people and build bridges across by eliminating barriers. AI upon threat concerns mitigation and if used productively can prove to be of tremendous use to the society and its people.

5. Can AI help the Government?

“ A global technology revolution is now underway. The world’s leading powers are racing to develop and deploy new technologies like artificial intelligence and quantum computing that could shape everything about our lives — from where we get energy, to how we do our jobs, to how wars are fought. We want America to maintain our scientific and technological edge, because it’s critical to us thriving in the 21st century economy. “ — ANTONY J. BLINKEN, SECRETARY OF STATE, from State.Gov

AI is already making rounds in the government sector. Examples include United States Postal Service (USPS) already using a handwriting recognition mail delivery system that uses AI as its technology. The United States Government has set up a website to bring its agencies under one umbrella namely — AI.gov and the above quote helps that the Government is considering AI optimistically but with caution against threats posed by AI. AI is also considered by other countries governments such as the National Health Service of the United Kingdom, which set up a chest imaging system for COVID-19 disease to detect and prevent the spread of disease among its people, and was used to support AI assisted technology as mentioned here. Not only the U.S. and the U.K., but also Singapore’s “Ask Jamie” virtual assistant helped its citizens and businesses to navigate across its close to 70 government agencies using AI powered chat and voice technology.

All of the above only suggests that AI is increasingly making its mark in the government sector and countries (both public and private domains) across the world are racing to tap AI’s power.

6. Threats posed by AI

Following are the concerns that needs to be addressed by any AI model implementation whether it is in public or private domain:

1. Privacy — Is the AI model considering privacy according to the law of the land?
2. Security — Is the AI model comprehensive enough to be protected against cyber security threats and risks?
3. Fairness — Is the AI model fair towards all segments of the people?
4. Transparency and Explainability — Can it be possible to demonstrate how an AI model operates and is it explainable?
5. Safety and Performance — Can desired safety and performance be achieved consistently? Is the AI model tested well enough to ensure the above?
6. Third-party Risks — Did the 3rd party vendors follow government regulations on risk mitigation?

7. Key Takeaways 🔑

1. AI’s capabilities are to mimic human senses and capabilities and if possible go beyond them.
2. Using a variety of subfields and components that AI is composed of, it is already revolutionalising itself in different industry domains such as Retail, Manufacturing, Telecommunications, Healthcare and so on with numerous applications.
3. AI is already proving beneficial to society and its people using a variety of applications and governments across the world are racing to tap the power of AI, with caution, due to the threats posed by the technology.
4. Indeed with every technology comes its risks and challenges but if one has the mitigation plan to overcome its risks and threats posed, certainly the technology will become a boon to its people, is what I strongly believe.

8. What’s Next? ⏭️

We will take a deep dive into an End-to-End Machine Learning LifeCycle where we not only look at the business objectives but portraying those business goals into a technical problem and solving them.

Don’t forget to give this article a 👏 if it sparks curiosity for the AI technology and the domain itself.

9. References

1. [Coursera] — Types of AI
2. [ScienceDirect] — AI in a selective industry
3. [Wikipedia] — Data Engineering