For quite some time people have been speaking about “Big Data”. With the advances in Cloud – both in terms of performance and price, it wasn’t surprising to see Google announce at Cloud Next in San Francisco that they were going to put AI and Machine Learning (ML) at the heart of every product.
Chief Data Scientist Fei-Fei Lee spoke of Google’s desire to democratise AI, so it’s easy and accessible to all. She shared a number of API’s allowing anyone to use Google’s Image, Text and Video Intelligence Machine Learning algorithms.
So what does AI and ML mean?
AI and ML is used to describe algorithms which allow a machine to reach a decision from its own logic. Traditionally, this could have been achieved by programming complex rules or logic statements to say “If this happens, then do this.” ML goes much further than this…
ML is designed to model how humans learn. It’s modelled with a series of inputs, connected to one or more “hidden layers” connected to an output layer. Based on the input; functions in the hidden layers transpose this into an output. Let’s say you wanted a machine to say if a picture contained a bus or not. You would first need a training set of images, some with buses, some without. Then you would feed each image in turn to the machine and it would tell you if there was a bus or not. If the answer was incorrect, this would be fed back and the functions within the hidden layer would be tweaked slightly before the next image is presented. Thus making it better-equipped to perform the task to a higher standard in future.
Why the buzz now?
Well, as you can imagine, the complex programming meant that you needed a strong understanding of maths in order to produce a Machine Learning algorithm. Cloud computing meant the power required to train networks at scale was available and you only need to pay for what you use. As Google has democratised AI, they have made TensorFlow their machine learning library open source. These functions are now freely available to all. On top of that, some of the models which do tasks such as image recognition, text to speech and sentiment analysis have been simplified as APIs. This allows you to simply pass an input in and receive an output without training.
It’s also important to note that these are not only unique to Google, Microsoft, Amazon and IBM. They all have offerings; meaning it is easier than ever to integrate ML into an application. In layman’s terms, AI and ML are much more manageable and accessible for everyone.
How does this benefit HR?
Employee wellbeing is another hot topic at the moment. Measuring and understanding staff feedback and insight is a key component to helping businesses create an engaged and productive workforce. Tools like The Happiness Index use pulse surveys and allow you to measure staff happiness, plot trends and use taxonomies to highlight key areas of concern. ML can allow this to be taken one step further. It is possible to train a Machine to identify key behaviours of “at risk”/unhappy employees and score this into low, medium or high. Using this data would allow businesses to proactively work to retain staff before issues escalate beyond resolution.
Other techniques such as sentiment analysis can be used to help pull out negative themes, and text mining can be used to extract key themes automatically. Image and speech recognition can also make a big impact. Amazingly, a machine could be trained to understand facial expressions. This can be used to understand how employees respond to particular meetings or situations. Equally speech recognition could be trained to listen for signs of stress or anxiety. Another use could be to recognise and allow staff to give feedback to Google Assistant, Siri, Alexa or other voice devices.
What are we doing with AI & ML?
At The Happiness Index, we have begun looking at how AI and Machine Learning can be further integrated into our platform and used to help unlock insight for our customers. We believe the possibilities and advances within this space will only continue to develop and we look forward to embracing these and other advances as they happen.