TensorFlow – the Google machine learning technology

4 Minute Read
Blog banner

Google’s recent innovation opens up machine learning to everyone, but what are the benefits of this and how does it affect search?

On Monday November 9, Google announced that it was developing smarter machine learning with TensorFlow. In its usual excited-five-year-old kinda way, the Californian firm explained how this technology was changing the face of search, both inside and outside of its own ubiquitous search engine.

What is TensorFlow?

TensorFlowTensorFlow is a new machine learning system created by the geniuses at Google. It’s been made scalable, so can be used over thousands of machines, from handheld devices through to huge datacentres. If you use ‘okay Google’ functionality then you’ve already experienced TensorFlow.

Data flow graphs are used to create connections between complex pieces of information, which can then be applied to seemingly complicated tasks such as speech recognition. A big advantage to TensorFlow is its flexibility; it can be run in CPUs or GPUs (the ‘brain’ in most computers) meaning that it can be deployed across thousands of different machine types.

Up until now Google has been using this software internally, but has now decided to make it open-source. In doing so, it hopes to encourage more use of this technology, to explore the plethora of amazing things that Google’s experts think TensorFlow can achieve. It’ll allow everyone from hobbyists to researchers to use the system for hundreds of other purposes. In doing so, Google will no doubt discover even more possible applications.

What does it mean for search?

We in digital marketing should be hailing this new opportunity and embracing it with open arms. Machine learning is a new and exciting field which will impact many industries in the near future, but this is something we can be utilising today. We’re already working on using it in our Apollo Insights platform to improve how we analyse websites and help us make accurate, informed decisions.

Automation of tasks

If I give you thirty seconds to write down all the mundane things you have to do with data, day to day, I’m willing to bet you could easily come up with at least ten. Working in SEO means the world revolves around data and I could easily list ten things that I could save hours on if a machine could intelligently do this for me.

In fact, our platform, Apollo, is already doing this.

Imagine if you took out all those mundane tasks that are super simple but incredibly time-consuming. What could you do with, say, one extra hour a day? How about two? What about four extra hours a day? How much more productive could you be? Automation is one of the most exciting prospects of machine learning. If a machine can take over those mundane tasks, we would be free to get on with the more important ones; applying the knowledge that those tasks provide us with – driving results much more rapidly.

Identifying trends

Machine learning will allow the rapid and constant identification of trends, based on real-time data. It could help you identify patterns that lead to revenue-generating ideas within seconds rather than hours or days. At the press of a button, you can take a massive amount of information, have it processed, and come out at the other end with a set of actions or solutions which you could apply to your carefully-crafted strategy or action plan. We all look for trends and patterns in all corners of marketing. A machine that can learn to identify patterns based on hundreds or thousands of factors will give us as marketers a never-before-seen level of data power. That’s another exciting prospect.

Making intelligent forecasts and predictions

You know those forecasts you did for the next financial quarter? Or those visibility and traffic graphs you put together months ago that took you nearly as long to work out? This could all become a thing of the past and, on top of this, they could be really accurate. In years to come (or probably sooner) you could be making accurate predictions, based on intelligent criteria, that are more likely than ever to come true, based on solutions that you and the power of machine learning have collectively driven. We’re not talking 30 or 40 per cent here, we’re talking five per cent +/- or even more so. This is just another incredibly exciting aspect of the sorts of things Google’s TensorFlow is leading towards.

This is just the beginning

Machine learning is an exciting and constantly evolving technology. This isn’t the first or last time we’ll hear about progress in this area. What it means to digital marketing and search overall is also ever-evolving. We have more data than ever and that data pool will only continue to grow. Machine learning systems like TensorFlow will help us use and manage it more effectively and intelligently. We have to evolve too, learning how to use this technology and put it to good use. It’s something that cannot be ignored and needs to be embraced. Through creating automation solutions and developing strategies there are hundreds of new and exciting ways in which the marketing industry can use this technology.

Check out this related article

Big data, algorithms, machine-learning… it’s happening around us as we speak and those companies that embrace it now will be the ones that reap the huge rewards: The big data future is here – are you investing?

Dave Colgate profile picture
Dave Colgate

Dave is head of SEO at Vertical Leap. He joined in 2010 as an SEO specialist and prior to that worked with international companies delivering successful search marketing campaigns. Dave works with many of our largest customers spanning many household names and global brands such as P&O Cruises and Harvester. Outside of work, Dave previously spent many years providing charity work as a Sergeant under the Royal Air Force Reserves in the Air Cadets sharing his passion for aviation with young minds. He can often be found in the skies above the south coast enjoying his private pilot licence.

More articles by Dave
Related articles