The first thing that comes to my mind when someone says AI or machine learning is terminator (The movie),
via GIPHY
Well this is not what we refer to as "MACHINE LEARNING" today at least not in the year 2018. If you are the everyday person trying understand what this hype is all about this may be the post for you.
Well with the reason developments I believe it is possible to create proficient AI. AI could be the next paradigm shifts alongside tech like quantum computers and so on then again there are downsides too like this.
Jerry Kurata. (2016). Getting Started in Machine Learning. [Online Video]. 17 May 2016. Available from: https://www.pluralsight.com/courses/python-understanding-machine-learning. [Accessed: 25 July 2018].
OxfordSparks. (2017). What is Machine Learning?. [Online Video]. 11 January 2017. Available from: https://www.youtube.com/watch?v=f_uwKZIAeM0. [Accessed: 26 July 2018].
Andrew Ng | Stanford University | Coursera. (2018). What is Machine Learning?. [Online Video]. Unkown. Available from: https://www.coursera.org/lecture/machine-learning/what-is-machine-learning-Ujm7v. [Accessed: 26 July 2018].
Well this is not what we refer to as "MACHINE LEARNING" today at least not in the year 2018. If you are the everyday person trying understand what this hype is all about this may be the post for you.
Machine Learning layman's terms....
Let's take a good example, Humans "us". We are capable of learning new things. From the moment we are born, every human start learning, discovering and completing tasks to achieve their day to day activities.
What if I told you Machines are capable of the same process too. With a strong combination of both statistics and computer science, machine learning researchers or computer scientists have proved that the same process could be adapted by machines too.
via GIPHY
How does statistics and computer science combined help machines learning. Let's understand this by a simple example.
For an instance, In grade school there is some sort of grading scheme for our exams some of which would be based on marks, subjects or other factors if we gave the data(Student Details, Marks & Subject) of a particular class in the school to the machine or computer. We would be able to utilise a set of algorithms to understand the most accurate approach to predict the grades of the students in the class.
What happens here is the machine tries to learn the pattern in which the grades are scored based on the data provided and produces a results based on the pattern in which it was trained. This is known as predictive analytics. (Not 100 % accurate ! but close enough to freak you out)
via GIPHY
If you are...... there are a few examples below that would help one understand.
What if I told you Machines are capable of the same process too. With a strong combination of both statistics and computer science, machine learning researchers or computer scientists have proved that the same process could be adapted by machines too.
How does statistics and computer science combined help machines learning. Let's understand this by a simple example.
For an instance, In grade school there is some sort of grading scheme for our exams some of which would be based on marks, subjects or other factors if we gave the data(Student Details, Marks & Subject) of a particular class in the school to the machine or computer. We would be able to utilise a set of algorithms to understand the most accurate approach to predict the grades of the students in the class.
What happens here is the machine tries to learn the pattern in which the grades are scored based on the data provided and produces a results based on the pattern in which it was trained. This is known as predictive analytics. (Not 100 % accurate ! but close enough to freak you out)
If you are...... there are a few examples below that would help one understand.
Experts Definition....
Machine Learning: Field of study that gives computers the ability to learn without being explicitly programmed. - Arthur Sammuel
Well posed Learning Problem: A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P, improves with experience E. -Tom M. Mitchell
A timeline of Machine Learning....
How,Why the hype & Why now ?
This particular hype started with Apple unveiling speech recognition on mobiles in 2008 which led to few more incidents IBM Watson defeating the best Jeopardy player in 2011 and more. Now with all these software advances came in high performance GPUs (A hardware upgrade) for gaming. These GPUs opened up new door for real life Machine Learning or AI implementations. Deep Learning techniques were applied and developed.
Today with the development of technology people are trying to solve the problems faced by our previous generations. Starting from predicting climatic changes, medical assistance to forecasting profits a business all these could be done with the help of machine learning.
Since this has become a game changer across industries most tech giants like Google, Facebook and Microsoft are investing by the millions to get there first prove that they have got what it takes to create artificial intelligence. While it seems like a long shot most of the have proven it with real life applications.
Do you use AI or Machine Learning daily ?
If you are not aware let me make this one clear, YES you are utilising some sort of AI or Machine Learning driven application.If you wondered how Google provides you with the best search results when compared to other search engines it utilises system called RankBrain this happeneds to be one of the many unknown number of strong Machine Learning systems used by Google Search.
Once you do a search on Google you suddenly start seeing advertisements maybe of particular product. CREEPY? Yes to an extent but this is marketing technique basically your search history is shared with Google's ad platform to make sure you see what wanted earlier. Again this is machine learning in play.
Another creepy feature by a tech giant is the auto tag feature on facebook. Just like in the above picture facebook over the years learns the facial features of an individuals by training a tagged picture with the particular individual.
If I were too list down example this post would be really long......
Should you be worried ?
Well personally I don't see something that should frighten you right now but maybe in the next 5-10 years yes. Given the data we are feeding public platforms I believe most of these tech giants will have enough data to clone a perfect individual or more like AI ?
Conclusion - overhype : not ?
References
Jerry Kurata. (2016). Getting Started in Machine Learning. [Online Video]. 17 May 2016. Available from: https://www.pluralsight.com/courses/python-understanding-machine-learning. [Accessed: 25 July 2018].
OxfordSparks. (2017). What is Machine Learning?. [Online Video]. 11 January 2017. Available from: https://www.youtube.com/watch?v=f_uwKZIAeM0. [Accessed: 26 July 2018].
Andrew Ng | Stanford University | Coursera. (2018). What is Machine Learning?. [Online Video]. Unkown. Available from: https://www.coursera.org/lecture/machine-learning/what-is-machine-learning-Ujm7v. [Accessed: 26 July 2018].