Why should anyone care about artificial intelligence?
I have always been fascinated about technologies and Artificial intelligence (AI) is one of those key technologies that could be defining our kids lifetime and ours.
In many research reports from around the world, there seems to an underlying current that automation is coming and it will be replacing our jobs. Combining the effect of technology growth and COVID-19, the pandemic of this decade, perhaps century, AI is rapidly bringing unprecedented capabilities to software, machines and robots.
This worries me as a parent, as I ponder upon the livelihood of the next generation if everything is going to be automated away. The way I deal with this worry is to find out more, to figure out what I do not know so that I can prepare for it.
Is AI new?
This is by no means a new field of studies, that just pop up in this decade of Computer Science. In the early days of Modern AI, there is the turing test proposed by Alan Turing for machine intelligence and even 2 major AI winters. One in the late 1960s to 1980s and then another in the late 1980s to 1990s.
Deep Learning
For what it's worth, Deep Learning has been considered the revolution of the decade. Starting from Li Fei-Fei's creation of ImageNet in 2009, to AlexNet in 2012, the improvements have not stop. There have been more than 2000 papers annually on "Artificial Intelligence" and only increasing.
Reinforcement Learning
From MIT Tech Review's analysis, neural networks are not the only one gaining in strength but there is also another domain called reinforcement learning. Beginning with Rich Sutton et al's work in the 1984, to David Silver's seminal paper on AlphaGo in 2016, reinforcement learning is growing rapidly as well.
Trends
What has caused this acceleration? Most will suggest the following 2 trends that are pushing this forward, Moore's Law for compute power and the accelerated generation of users' data. With more than 3.7 billion people using software on their computers and mobiles, we are seeing 2.5 Quintilian bytes of data created each day.
Even as we start to see a gradual decline in Moore's prediction of the growth of transistors on the chips, newer technologies such as quantum computing or the rapidly improving GPU compute are changing how much AI computations can a researcher or a research lab do.
Access to Knowledge
And I will argue the 3rd trend that is accelerating all this development is the access to knowledge. With COVID-19, we started seeing more open access provided to plenty of knowledge online. In fact, we are now faced with a new problem, the "information overload" era.
Robotics
After learning about AI - Machine Learning, Deep Learning and Reinforcement Learning topics, frankly, I still do not know much. Papers submitted on Robotics grew by thirty-fold between 2010-2019. But how do all these affect Robotics? Well you might have guessed it, jobs displacement is once again the problem.
Robots May Displace 20 Million Manufacturing Jobs by 2030 - Bloomberg
Machine vision has traditionally been done with handcrafted features but Deep Learning changed many of these domains by using an automated feature generation process or Convolutional Neural Networks (CNN).
Learning as a Newbie
I started learning about all these via MOOCs as my main source of information. But the amount of good information out online is seemingly endless, you might not know where to start or stop.
Going forward, I will be sharing my Trello board that lists the MOOCs/resources I am looking at as well as the upcoming promising fields of research interesting to Robotics and Computer Vision domains.
Start today
If you managed to read this far, I believe you agree with me on the importance of AI in the future work and society of our generation and the future generations to come. So start today, learn more about AI.
AI Newbie
AI Newbie will be my blog for documenting business, technical know-how and knowledge on AI that I find important, and useful for my future references. I hope it will be useful to anyone out there reading as well.