AI has learnt a lot from humans; can we learn anything from AI?
This article is a food for thought on one key learning humans can adopt from Artificial Intelligence. I encourage anyone and everyone who reads this article to express their opinion on this subject. In last 5 decades human intelligence have marginally grown but AI has achieved exceptional heights starting from almost nothing. Thus, evolution of AI is much steeper than human's. So there is something right which AI are doing which we can adopt. What is fundamentally different between the way human's learn and the way AI learns. Can we use this difference to increase the efficiency of our education system? As the maximum learning human does is in the early age, the majority of the changes which will come out of this comparison will be in early schooling. Obviously, this section is just my personal opinion and might conflict with many.
Have you entered a classroom where the professor/teacher started the class with "Here are the things we know about this subject and here are the things we do not." I am sure majority of you will say "No".
Here is my personal experience over time
At an age of 15, I was introduced to "Newton's Law of Gravitation", but I was not told that this law is just over simplification of gravity and does not explain where does gravity really come from. All I knew was the area inside blue circle and I had no clue that there are valid theories beyond blue boundary. Because I was told that this is the complete theory I need to know, I never even wondered on anything beyond. At age of 19, I got to know Einstein was able to explain gravity accurately and there is nothing beyond green boundary. I again did not think beyond the green boundary as I did not even know that there was anything beyond. At age of 22, I again got to know that this theory of Einstein does not align with Quantum mechanics and again that is all I need to know. I did not even evolve my ability to access this lie with time because they happened with big intervals of time.
Obviously, you do not want to scare a 15 year old kid by showing him Quantum mechanics theory, but we need to make them aware that there is world beyond this current boundary. I will talk about this knowledge as "things we know we don't know" or "the known unknowns". This might be sounding philosophical but let me elaborate with a professional experience of my own.
I was recently given the task of using Machine Learning to optimize Citi's IVR (Interactive Voice Response). Given this is a new field of research, I did not find any good documentation of such project across industry. So I had a "known unknown" and started my exploratory journey on all possible solutions. I eventually landed to 5 possible fixes out of which 4 are now a part Citi's IVR leading to huge expense savings. I, then, attended a conference called SpeechTek where I realized only 2 of these 4 feasible solutions are being leveraged across industry. Now I realize if I was a part of SpeechTek before finding the 5 fixes, I would have restricted the scope of my search to the known 2 solutions. The exploratory part enabled me to think outside the box!
No knowledge of boundaries of the box leads you to thinking outside the box.
I am sure many of you would have had this "Eureka" moment where you did an exploratory analysis which led you to think beyond what industry has achieved. So why not include this learning to the stage where human's learn the most - early schooling. Why not give more open syllabus to learn from rather than closed bounded circles.
So, who does this kind of learning flawlessly?
Artificial Intelligence!!! Have you interacted with Alexa/Siri and asked a question out of syllabus? What answer did you get? Probably something like "I currently do not have the capability to do this task." So AI exactly knows what it does not know - "the known unknowns". AI then learns from these new data points to explore this unknown land.
Handful of countries have adapted this methodology to build education framework. However, majority of developing countries still believe in defining the exact scope of courses in early schooling and even in higher education like bachelors and masters. I believe Artificial Intelligence provides us a strong analogy of how to evolve our education system. I will love to know your opinion on this ideology and what other things can we adapt from the evolution of Artificial Intelligence.
AI | Product Management | Program Management
7 年https://www.dhirubhai.net/feed/update/urn:li:activity:6370659942457753600
Open to New Roles | LGBTQIA + Network Leader
7 年Massive amounts of data is coming in everyday and many companies are working on understanding what is important from what is not. However, this data still needs to be stored somewhere? Generally, what I have heard from many companies are the endless crashes or bottlenecks they go through in Machine Learning? Analytics? The list is long. My question is, given what I expressed above, would this happen to AI driven computers? Will, they have the ability not to crash or slow down , Bottleneck etc?