Are You Ready for the Next Technology Revolution?
Imagine launching a product in 2010 that within just 5 years had sales of over 250 million and has become a must have for individuals and businesses alike. Impressive huh? Apple’s iPad has achieved just that, and helped many follow products and spin offs succeed along the way too. Without doubt it has truly revolutionised our domestic and working processes in ways we could not have imagined.
As one of the early iPad adopters I remember well the doubting looks and snide comments from some who saw it as a gimmicky gadget. Now those same looks and comments come from the same people towards those that don’t have one!
In so many ways the admirable iPad has helped us speed up those things that we have always done allowing us to do more of them “any time, any place, any where” – and with the extra time we’ve created for ourselves…… Hmmm – we work harder! For many work life balance is worse not better, not what we may have predicted.
So looking forward what could the next 5 years hold – more of the same or is it time for another technology revolution? Well my money, literally, is on the latter and whilst between jobs I’ve taken time to understand Artificial Intelligence (AI) Deep Learning and its potential. Consequently I am investing time and energy in backing this as a winner of the future. If you want to find out more about it and its potential then read on.
Deep Learning is about software bots, (use this link, survive the ad and be taken directly to the video extract point to learn more about bots – https://www.youtube.com/watch?v=7Pq-S557XQU&t=7m25s), that teach themselves to perform complex tasks. Some tasks are now too complex for a programmer to be able to write the instructions to perform them properly, so it is better left to the machines to do it themselves. For a close to home example take motor collision damage assessment: the best motor engineers draw on many years of experience and we respect them and their work as an art. The rules they work with are very subtle. Deep Learning software bots can learn these subtle rules by feeding on images of motor collisions tagged with the appropriate repair estimate metadata (metadata for those like me who did not know is simply data about data).
Feeding an image into the software is just like a teacher pointing to a damaged car and explaining to the student what repairs are required. But here’s the difference: an image engineer with many years of experience has learned from several thousands of cases; a software bot can learn from 10,000 cases in just a few hours. For sure that is scary, yet also an inevitable development of the future.
We are approaching the point where AI researchers can pioneer the exact learning environment that’s required for the software bots to quickly overtake human capabilities, perhaps even starting to predict things that humans can’t gather enough experience to be adequately informed about. That is exactly what Tractable, the organisation I am working with, has set out to do. This video shows the full potential of the technology. It’s well worth a few minutes of your time.
So, keeping close to home, what if Deep Learning came to motor repair estimating? What if these software bots were to be deployed on cameras, smartphones and the Internet of Things? What if this visual damage assessment were combined with the internal vehicle data gathered by telematics devices? Could the entire motor claim pipeline come under the control of those who command these technologies? That’s the bet I’m making – and the pace at which the simple iPad has changed our world suggests this could happen quicker than we all think.
But it doesn’t stop there: algorithmic vehicle damage assessment could be used to assign liability from CCTV footage of accidents, to assess car hires before and after hire, to match light repair private jobs to a marketplace of repair shops, to allow manufacturers to track their vehicles throughout their supply chain, to create a peer to peer platform for selling used cars… and the list goes on. Thanks to my 25+ years of experience in this industry, I’m making sure Tractable finds the best partners for their technology. Interested?
Of course, that’s just vehicle damage assessment. Any other visual recognition task can be automated by Tractable’s technology, and they’re on the hunt for more datasets to feed their bots…
The concept is simple, clear and there’s an inevitability about it coming to a place near you soon. Some of the biggest technology companies “in the know” are buying up Deep Learning expertise so they can capitalise on the potential. It’s creating quite a buzz in the next generation technology world.
Taking the concept into business reality is currently the missing link – which is why I am working closely with a great group of talented guys at Tractable who want to see this through and won’t be looking to sell out for fast buck. Developing this AI requires world class talent, and that is exactly what Tractable’s team of 6 AI engineers boasts. The CTO, Razvan Ranca, was at the Machine Learning Group in Cambridge where he worked on probabilistic programming, while the CEO, Alex Dalyac, worked on industrial applications of AI while at Imperial College London. With around US $2m of investor support I’m not alone in thinking these guys are the winners of tomorrow and their success will see the disruption of markets in ways we can not yet imagine.
In a short period of time they have identified many activity streams in a variety of industries where bulk manual inspection by technical experts can be replaced by automation giving far higher quality at far lower cost with a significant reduction in process life cycles and a potentially impressive downstream uplift in customer service. This is a win win for all but those who wish to sustain the status quo of yesteryear – early iPad naysayers no doubt!
One of those initial areas of investigation, the inspection of motor damage repairs, has seen impressive progress towards proving the concept – exceeding initial accuracy targets by some way in considerably shorter lead times than expected. Things are really gathering pace so watch out!
However there’s a word of caution to fellow insurance market Brits. Our US insurer colleagues, the motor manufacturers and other players of the future insurance market are far more embracing of the potential. The traditional UK market with its short term focus, steady pace of change, vested interests and quirky, unsustainable ways of generating extra revenue from claims incidents, runs the risk of being left behind.
As the unstoppable juggernaut of Deep Learning comes over the hill industry leaders should quickly take note. This next generation technology really will help us work smarter not harder and those that don’t embrace it could well find that work life balance improving sooner than they had planned.
If you wish to find out more about this topic feel free to contact me via LinkedIn or email@example.com, or Tractable’s CEO Alex Dalyac via LinkedIn or firstname.lastname@example.org .
Posted 18 August 2015
Source: Are You Ready for the Next Technology Revolution?
Via: Google Alert for Deep Learning