Frequently Asked Questions about AI
It is of no doubt that the field of Artificial Intelligence has created a big hype around it in the last five years. The field has done wonders in awakening the curiosity and the interest of most of the population on our planet. Its multidisciplinary nature, as well as the promises it makes to transform careers, are some of the factors that made most people interested in understanding this technology regardless of their technical background or age.
As an expert and professor in the field, I wanted to contribute to spreading the knowledge about AI by answering a set of questions that are frequently asked by curious AI novices.
I hope you find this FAQ useful! If you have any other questions, don’t hesitate to ask! I am also open to debate and discussions :).
What (the hell) is Machine Learning? And how it differs from Artificial Intelligence?
Before defining Machine Learning and Artificial Intelligence, I think it is essential to define intelligence. Intelligence can be defined as one’s capacity for logic, understanding, self-awareness, learning, emotional knowledge, planning, creativity, and problem-solving. It can be more generally described as the ability or inclination to perceive or deduce information and to retain it as knowledge to be applied towards adaptive behaviors within an environment or context.
Artificial Intelligence is the broader concept of endowing machines with intelligence, whether it is emotional intelligence, social intelligence, logical intelligence, planning, creativity, etc.
The birth of Artificial Intelligence as we know it today dates back to 1956 where researchers from Dartmouth came together with the explicit goal of programming computers to behave like humans.
Machine Learning is a subfield of AI, it can be seen as a way to implement decision-making in AI and getting computers to learn. You probably have heard of deep learning which is a subset of Machine Learning and the most effective of all the machine learning algorithms.
Why did it became so popular?
One major breakthrough that led to the emergence of Machine Learning as the driving force behind Artificial Intelligence is the invention of the internet. The internet came with a huge amount of digital information being generated, stored, and made available for analysis. This is when you start hearing about Big Data. Combined with the advances of computing power, Machine Learning algorithms have been the most effective at leveraging all of this Big Data.
Machine Learning is popular because ML algorithms actually work. Object recognition algorithms have scored a performance equal to the performance of humans. ML nowadays is being applied to almost all industries, and soon, many companies that ignore ML will be crushed by new startups that embrace it.
I heard of strong and weak artificial intelligence — what it is?
Weak Artificial Intelligence, also known as narrow AI, is AI that is focused on one narrow task or problem. All currently existing systems that are considered as Artificial Intelligence of any sort are weak AI at most.
Siri is a good example of narrow or weak AI. Siri operates within a limited predefined range, there is no genuine intelligence, no self-awareness. If you ask questions outside the limits of the application, don’t expect Siri to respond hastily.
Strong Artificial Intelligence is a machine that has consciousness, sentience, and mind. An Artificial General Intelligence is a hypothetical machine that is capable of applying intelligence to any problem, rather than just one specific problem. Such a machine is assumed to manifest behavior at least as skillful and flexible as humans do. The machine can actually think and perform tasks on its own just like a human being.
Personally, I am a strong supporter of the computational theory of mind, the philosophical position that human minds are, in essence, computer programs. Such a position can be named strong AI.
I strongly believe that if we reach a point where we truly understand the underlying mechanisms of the human mind, we will be for sure able to produce a strong AI machine. However, I certainly don’t believe that a machine can evolve by itself to become more intelligent than how it was initially trained. This means that a machine can never surpass itself. In my opinion, this can only happen if first as human we are capable of surpassing our brain physiological capabilities; two if we understand how our brains evolved to be able to do that; and three if implement this mechanism in AI algorithms and systems.
Explain the most interesting example of using machine learning/artificial intelligence that you ever heard of.
One of the most interesting examples of using ML is in my point of view is human behavior and mental state automatic understanding. This field is called social Artificial Intelligence and it concerns building machines that are socially and emotionally intelligent. What’s fascinating about this field is that it is highly multidisciplinary, combining findings from psychology and social sciences with computational sciences. If you are interested in understanding more, check out some of my publications on the subject here. I have published research on personality computation, emotion recognition, and engagement detection.
We hear that Machine Learning and Artificial Intelligence are somewhere near us. Where can we find them in everyday life?
There are many examples of AI and Machine Learning at work in our world today, that touch our everyday lives, but some aren’t even aware of it. For example, every time you do a web search, when Netflix recommends a movie when Facebook selects posts, when Amazon recommends a book, it’s machine learning that’s behind. Other applications of Machine Learning or AI exist in domains like robotics, vision, and natural language processing, or medicine, oceanography, social science, you name it.
I would like to learn more about it. Where can I find some useful informations and tutorials?
A great book about AI which can be considered as a reference is “Artificial Intelligence: A Modern Approach” by Stuart Russell and Peter Norvig.
This book offers a complete introduction to the theory and practice of Artificial Intelligence. This book can serve as a great reference for undergraduate or graduate-level courses in AI. The authors offer a free online course at Stanford University on AI.
Other courses exist online on Coursera.
Machine Learning Specialization from University of Washington
A four-course specialization that will teach you everything you need to know about AI. The specialization starts with teaching you the foundations of Machine Learning followed by deep diving courses. You’ll need to commit to a few months of study, with each course running at least 6 weeks.
The foundational course alone provides practical, real-world experience in applying machine learning, so it’s a great starting point if you’re unsure about taking on the other courses too. Projects you’ll encounter in this specialization include predictive analysis, classification, and clustering of information using machine learning algorithms. Find details on each course here.
Deep Learning specialization from deeplearning.ai, taught by Andrew Ng
If you want to break into AI, this Specialization will help you do so. Deep Learning is one of the most highly sought after skills in tech. This course will set you on the track to become good at Deep Learning.
The specialization includes five courses, where you will learn the foundations of Deep Learning, understand how to build neural networks, and learn how to lead successful machine learning projects. Projects you will work on include healthcare, autonomous driving, sign language reading, music generation, and natural language processing.
This course can help you build a career in AI. You will master the theory of AI as well as its application in various industries, all in Python and in TensorFlow.
Data Science -> You said Data Science? [explain what it is]
The term “data science”, originally used interchangeably with “datalogy”, has existed for over thirty years and was used initially as a substitute for Computer Science by Peter Naur in 1960. It, is an interdisciplinary field of scientific methods, processes, and systems to extract knowledge or insights from data in various forms, either structured or unstructured, similar to data mining. It includes data cleansing, preparation, and analysis.
According to Harvard Business Review Data Science is “The Sexiest Job of the 21st Century”, so you better get started right NOW!
What is a Data scientist?
A Data Scientist gathers data from multiple sources and applies machine learning and predictive analytics to extract critical information from the collected data sets. The objective of a Data Scientist is to understand data and to provide accurate predictions and insights that can be used to power critical business decisions for example.
A data scientist should have the following skills:
- Strong knowledge of a programming language such as Python, SAS, R, or Scala
- Hands-on experience in database coding such as SQL
- Ability to work with unstructured data from various sources like video and social media
- Knowledge of Machine Learning
How it differs from ML and AI?
As I said earlier, Data Science is about extracting knowledge or insights from data in various forms.
The word “Learning” in Machine Learning intends that ML algorithms learn from some data, used as a training set, to train and thus fine-tune some model or algorithm parameters.
AI makes use of Data Science and Machine Learning to develop machine’s “Intelligence”.
Machine Learning can be seen as the link between Data Science and Artificial Intelligence.
What kind of advice can you give someone who want to start his journey with ML/AI/Data Science?
Learning ML/AI/DS can be confusing, especially for someone who is just starting their learning journey. Many questions come to mind. Which tool to learn? What techniques to focus on? How much statistics to learn? Do I need to learn coding? These are some of the many questions you need to answer as part of your journey.
1. Choose the right role
There are a lot of varied roles in ML/AI/DS. Examples include data visualization expert, machine learning expert, data scientist, ML engineer, robotics engineer, AI researcher, etc. So before starting, do your research on what roles are there and choose the role that suits you the best. If after your research, you’re more confused than before, I advice you to choose the data scientist role. If you’re a data scientist, you won’t struggle much to find a job in any industry you like!
2. Start a course and complete it!!
After choosing a role, you should do what it takes to understand that role. By understanding the role, I mean that you should understand the requiremens of the role, search for relevant courses, start a course and actually completing it!
Completing a course is the most difficult part of this step. I can’t tell you the number of times I have started an online course, payed for it and did not complete it. To be honest, I am actually registered in a coursera specialization right now and also PAYING FOR IT since 8 months. Don’t do like me!!!
The key to completing a course is keeping in mind your objective and the benefits you’ll be getting when you finish it! YOU DO IT!!!
3. Choose a Tool/Language and stick to it
A difficult question which one faces in getting hands-on is which language/tool should I choose?
This would probably be the most asked question by beginners. The most straight-forward answer would be to choose any of the existing mainstream tool/languages. These include SAS, R and Python. Here is a nice guide comparing these tools.
However, keep in mind that tools are just means for implementation, and you can never go anywhere if you did not understand the concept!
My advice is to start with the tool that you feel the most comfortable with. This also depends on your technical background and if you are familiar with programming or not. If you are not that familiar with coding, GUI based tools such as Weka might be a good solution for you to start with.
