Head in the cloud?
How cloud services are bringing a new age of accessible technology.
The cloud has rapidly evolved in recent years and in turn has been adopted en masse by businesses, individuals and governmental bodies alike. When Amazon developed one of the very first cloud platforms in 2006, few would have believed the idea would have taken the world by storm. However, today the cloud is now an integral component of the tech industry.
This new and improved way of dealing with big data not only comes with the obvious benefit of off-site storage, but has also revolutionised the way we work within technology. Managed services available in the cloud are now allowing engineers, scientists and researchers from a wide array of disciplines and backgrounds to explore and experiment with technologies that would have previously been out of their reach.
At Ocado, the cloud has dramatically shaped how we do business. As early implementers of AWS and the Google Cloud Platform, we went from storing nothing in the public cloud, to storing over a petabyte of data in just 18 months! For more about our journey into the cloud, see How Ocado moved its databases to the cloud, Taking customer analytics to the cloud and Ocado’s journey to the cloud using Amazon AWS.
Our move to the cloud has had countless associated benefits; we have seen many exciting ideas related to our Ocado Smart Platform (OSP) product become a reality; we’ve also used the cloud to improve upon our customer service by using TensorFlow, Google Cloud Machine Learning and the Cloud Natural Language API to categorize customer emails (see our blog) and, more generally, significantly increased our scalability and efficiency. However, most importantly when it comes to our dev culture, the shift has empowered our developers and opened us up to new possibilities. We can now see our ideas becoming a reality quickly, which in turn promotes a widespread culture of autonomy.
This development has not just changed the way we work at Ocado Technology, but has also had an impact across the industry, and as a result we have seen many interesting uses for machine learning and AI spread to a general audience (albeit some more useful than others!)
Here are some of our top picks:
The chatbot that can help you appeal your parking tickets
In 2015 Joshua Browder launched his DoNotPay chatbot, designed to help motorists appeal unfair parking fines. The teenager, who started the project after appealing multiple tickets on behalf of friends and family, hosts his chatbot on Heroku in the cloud.
The endeavour has so far won 160,000 out of 250,000 recorded cases, and he has since expanded the capability of the artificial lawyer to deal with housing disputes and asylum cases, working in partnership with homeless charities in the UK and beyond.
He has recently gone one step further, however, and is now opening up his platform so that anyone can create a chatbot for free.
“We realised no great work is done alone. Starting today, DoNotPay is opening up so that anyone can create legal bots for free (with no technical knowledge).” Joshua wrote on Medium.
This is a perfect example of the cloud making advanced technology accessible for those without technical backgrounds.
For another example, see how this Android developer designed, developed and deployed a chatbot entirely in the cloud.
Your own digital voice
Canadian startup Lyrebird has developed a software that can mimic any human voice after being provided with just one minute of audio!
The software, which is currently available in beta, allows you to generate sentences using your voice recording as seen in this video of Barack Obama:
One of their key goals is to help those who have lost their voices to a disease, by assisting them in regaining this important part of their identity. However, there are plenty of other applications, including personalised dialogues within video games and text to speech messages read in the voice of the sender. One obvious concern is that this technology could easily lend itself to fake news and identity fraud, but Lyrebird is determined to make sure the technology is not used for such purposes. In a privacy statement released on their website they claim that “your digital voice is yours. We are stewards of your voice, but you control its usage: no one can use it without your explicit consent.” However, as with any new technology we need to be responsible when putting it into practice.
The technology is still in the early stages; testing the beta version privately with selected developers. The digital voices are not perfect, but they are able to capture some of the nuances that make up an individual’s tone, and are therefore recognisable. They do clarify, that while they can create a digital fingerprint of your voice from a minute of audio, the longer the clip you provide, the more accurate your digital voice will be.
Algorithms to help you sort through your holiday snaps
If you use Google Photos to store all your favourite cat pictures, family memories, selfies or whatever else takes your fancy, you might be aware of some of their cool categorisation techniques. You can now search through your photos by location, objects, themes, colours… pretty much anything you like, within reason.
Using powerful machine learning models and image recognition technology, Google are able to recognise your location based on the landmarks around you, even when your location services are switched off. You can enter whatever term you like in the search bar, like temple, concert, food, pets, beach, Prague etc., and images that return a positive match with that word will be displayed on your screen. Pretty neat.
We ran an image using their testing service from one of our backdated blog posts to see how it performs.
The same machine learning technology that powers the Google Photos app is available to anyone via Google’s Cloud Vision API, meaning access to this technology is at everyone’s fingertips.
There are plenty of applications for image recognition, and of course, like most technologies, some don’t serve an obvious purpose, like what-dog.net, which lets you upload a photo of your pampered pooch and identifies its breed (although after a little office experimentation, we discovered it also allows you to upload your own picture so you can answer the all important question: ‘What dog do I look like’), or Not Hotdog, the app that can tell you whether your uploaded image is a hotdog, or not (to read more about the making of the app have a read of this Medium post)
Earlier this year, at the Cloud Next Conference, Google announced the release of Cloud Video Intelligence in beta; a service capable of discovering and categorizing information from video. This marks a significant step forward in machine learning, as videos have long presented a challenge. Why not test out it’s capabilities by uploading your own videos here?
This article has only scratched the surface of what’s out there – going through the product catalogues for the various cloud service providers can give you a better idea of what’s on offer. The world of technology is now more accessible than ever, with basic programming knowledge and a good idea, you can start creating your own apps, services and games. So what are you waiting for?
Holly Godwin, Technology Communications Assistant