Machine Learning: A Quick Overview

Machine Learning: A Quick Overview

If you have been following the social media lately, you must be aware of the ongoing repartee between two powerful men- Tesla and SpaceX (SPACEX) CEO Elon Musk and Facebook CEO Mark Zuckerberg over what Artificial Intelligence (AI) would really mean to the future of humankind. The whole ‘future of AI’ is a bit too far-fetched for our understanding. But, as L&D professionals there is another term ‘Machine Learning’ that has caught our attention from amongst all those discussions about AI, Big Data, Analytics etc.

So, what does Machine Learning mean?

By Definition

According to Wikipedia, “Machine learning is the sub field of computer science that, in the words of Arthur Samuel,gives “computers the ability to learn without being explicitly programmed.” Technically it is a type of AI, that allows programs to run predictive analysis even without the actual commands to do so.

The algorithms for Machine Learning can be of two types: supervised and unsupervised. While supervised learning builds on the instructions for input and output and feedbacks about the accuracy of prediction during training. This learning is then applied on new data.  Unsupervised learning does not require a desired outcome data to learn from, instead it uses an iterative approach known as deep learning (also known as deep neural learning or deep neural networking) to analysis data and draw conclusions.

The simplest example of how machine learning works, is Google AdSense, that gives you suggestions about restaurants, shares advertisement on web-pages etc. all based on your search patterns.

Exemplification

Now dwelling a bit deeper into the deep learning. How does a machine learn exactly? In L&D terms who trains whom? For starters, it is a data driven process. The learning is described in form of neural networks that replicates the way biological neural network (brain and the connected neurons- the signal transfer etc.)functions.

Let’s think in terms of how a machine recognizes images of animals accurately. It first learns from the existing data, identifies recurring patterns to that differentiate one from the other and uses this learning for image identification.

Here’s an image that explains it quite well.

Neural Networks

PC: Fortune.com

Facebook News Feed, utilizes machine learning where it uses personal data (pages we like, friends we constantly follow, external website links we usually access etc.) to generate a custom news feed for each user.

Machine learning is powerful indeed. The more it learns, the more it can perform. Machine learning has greatly influenced processes like image identification, recognition of voice commands, and in running advanced searches too. While Machine learning is still evolving, major corporations like Google, Microsoft, Facebook already have it in place. IBM too has IBM Watson Machine Learning (earlier known as IBM Predictive Analytics service).

Wikipedia shares an extensive list of applications for machine learning. Medical diagnosis, Internet fraud detection, information retrieval, advertising, speech and handwriting recognition, user behavior analysis, translation etc. to pick a few.

Implication on eLearning

AI and Machine Learning may have both positive as well as negative impacts. While there already are few open source tools available to streamline deep learning projects, setting up predictive analytics programs can be far from easy and requires specific expertise too.

When we talk about a cognitive learning model with which machines can be smarter, does it also impact eLearning design, development and delivery as we know it. It may be a bit early to say so, but there is a possibility that machines will replace humans in fully automated development and testing. However, machine learning still has a long way to go in terms of eLearning content design, though it sure can be assistive in data collection, predictive analysis of preferable learning models etc.

We at Knowzies have recently started thinking on how L&D can be benefited from this and how best we can use Machine Learning concepts while designing innovative eLearning solutions. In our future blogs you can expect to hear more on Machine Learning from our experts.

We also would like to know on what are your thoughts on Machine learning and its influence on eLearning? Do share your opinion.

1 Comment

Leave a Comment.