Is a Cat a Cat?

Two questions for you:

1. When is a cat a cat?
2. How do you describe a cat?                        

You may think these are very simple questions…I bet you’re thinking two ears, four legs, whiskers – yes? Well how does that differentiate a fox, mouse or rabbit?

As human beings our brains are hard wired to a certain visual system that helps us recognise objects easily. From an early age, after pointing at a cat saying ‘dog’ a few times, our parents help us to soon learn the differences. However recognizing a cat, is exactly what a machine struggles with.

Think about when you learned to read. Once you recognised the shapes of the letters, you soon were able to read any book from the Hungry Caterpillar to Harry Potter. When we are a child we build up this visual system to recognize all sorts of shapes, whether hand written, typed or in pictures. When you change the lens to a machine, the amount of information given to a computer is limited. The machine does not have the same level of visual system, and therefore struggles to learn at the same pace as humans – phew, the world is not turning into iRobot just yet!   

But we help machines learn every day without realizing. Listening to Timandra Harkness talk on Big Data at Science Museum Lates session this month, it brought to my attention that every time we see one of those ‘I Am Not A Robot’ tests, we are not only keeping the ninja bots that like to attack our computers at bay, but also to assist in the digitalization of books. Our human deciphering of the distorted words is part of the validation stage of a machine recognizing a word or phrase.

Through the interplay of both human and machine, we are getting even savvier when it comes to machine learning. It has brought us effective web search, self-driving cars and even beer that learns – (yes you should definitely lookup IntelligentX, the world’s first artificial intelligence brewery).

So, when will the need arise to become Learning and Development specialists for Machines?

Being in the field of Learning and Development that is the question I pose. As machine learning moves beyond the study of pattern recognition, we need to question how we improve the ability to teach a machine or question whether it will ever be possible to have no need for human intervention.                       

Whilst we eagerly await for futurologists to replicate “human learning machines” that we often see in science fiction films, our focus should be on what machines can do, a great deal faster and more efficiently than people – with the lens of how L&D can support behavioral change among the workforce.

With powerful data-driven machines at our finger trips, we can empower our workforce to make more meaningful decisions across the business, but not without supporting the behavioural shift. With Big Data comes a greater need for capability in problem solving and decision making to make use of those boffin machines, rather than keeping them as an untapped secret.

We are not seeing a future yet where the need for people is eliminated, but as people in a world of greater machine learning we will need to think about how we think, act and respond intelligently to data predictions and ask big questions about how our L&D strategies don’t just consider people, but the machines they work with. 

Maybe a machine can teach a machine?