By Anisha Mata
If you’ve recently used Twitter to get your news, asked Google Assistant to call you an Uber, or talked to any online customer service, chances are you’ve been in contact with Machine Learning in action, without even knowing about it.
Since in recent years, machine learning has become such a big part of how we interact with our technology, it’s important we understand what it is, how it works and its potential for the future.
What exactly is Machine Learning?
Machine learning is essentially the use of artificial intelligence by computer systems to learn from experience (patterns and logic) and perform or improve tasks without being directly programmed to do so.
Interesting fact about Machine Learning: It is used by banks and financial institutions to prevent fraud and to give insights about data in seconds.
This process starts with data and observations. For example, if an Amazon creates a list of recommended products for you, it does so by observing the goods you have already purchased and other factors influencing those purchases. The computer system looks for patterns in the data so that it can recommend you better products, that you are likely to buy.
The goal of Machine learning is to allow the computer to do all the heavy lifting and learn from its experience without any human intervention.
It tries to find correlation between different variables.
How does Machine Learning work?
So essentially, Machine learning is an Algorithm. An Algorithm is a set of instructions that are followed to perform certain calculations, process data or find patterns.
This Algorithm finds a correlation between input x and output y and uses that to improve certain tasks.
Example of a machine learning model based on linear regression.
For example: A skin cancer detecting algorithm created by Stanford Artificial Intelligence Laboratory and Stanford Medical School, that uses a database of 1,30,000 skin disease images, has been proven to be just as effective at detecting skin cancer as a team of 21 certified dermatologists.
Such applications of Machine learning can actually save lives and reduce the time and costs involved in detection of diseases.
Interesting Fact about Machine learning: This term “Machine Learning” was coined by computer gaming and artificial intelligence trailblazer, Arthur Samuel, way back in the year 1959.
Applications of Machine Learning
As evident from the previous examples, machine learning has a wide range of applications, in almost all industries. From agriculture to healthcare, it can find its application wherever a large quantity of data can be provided.
Let’s look at some interesting examples:
- The Deep Blue was a chess playing computer created by IBM, which was the first computer ever to be able to defeat a reigning world champion (Garry Kasparov) in a chess game and a chess match under regular time controls.
- Academic publisher Springer Nature has published a research book in 2019 titled “Lithium-Ion Batteries: A Machine-Generated Summary of Current Research”written completely generated using machine learning.
- Machine learning plays a major role in the development of self-driving cars
- Search engine optimization also works on the basic principles of machine learning
- And in his way, if we pay a little more attention to the technology that we use, we will find several examples of machine learning in our life.
Limitations of Machine Learning
Although machine learning has transformed several fields completely and will continue to do so, we must acknowledge its limitations. The limitations in machine learning mostly arise due to lack of suitable data, biases in data collection, privacy or improper algorithms.
Let’s see some examples of the limitations of machine learning:
- In 2016, Microsoft tested a chatbot for twitter, which after learning from existing tweets started posting with racist and sexist language. (This is an example of unsuitable data)
- In 2018, an Uber self-driving car failed to detect a pedestrian who was killed after collision.
- Machine learning is also an expensive technology. IBM’s Watson system which was expected to play a huge role in healthcare failed to give desired results even after billions of dollars of investment.
Thus, we can see that even though machine learning is a transformative technology that has he potential to transform our lives, it must be managed with the utmost care, keeping in mind the consequences that may occur from its use.
Graph used is from: https://docs.microsoft.com/en-us/azure/machine-learning/studio/algorithm-choice