Ten Innovative Uses Of Machine Learning In Business

Machine learning is the intelligence powering more and more of our must-have technological developments. The branch of artificial intelligence allows computers to learn from data and make decisions, simulating the function of the human brain. The technology is becoming part of our everyday lives, helping us identify patterns and make faster, more accurate decisions. 

For businesses, machine learning offers impressive potential to optimize operations. It is being used in every sector, and the possibilities are boundless. Customer and client satisfaction can be monitored and tailored accordingly. Market trends can be tracked and reacted to in real time. Large data sets can be analyzed and risk estimated and minimized. Most excitingly, machine learning can give businesses a competitive edge, opening up new revenue streams and allowing for cutting-edge innovations. 

From your favourite online retailers to your medical providers, you have no doubt come into contact with machine learning. You may not notice it, It is often under the radar and by its very nature should fit seamlessly into everyday life. To shine a light on machine learning, we’ve picked out ten of the most innovative ways it is being used across various industries.

1. Accurate searches on Google Maps

Google is one of the technology giants at the forefront of machine learning and artificial intelligence. Last year they introduced machine learning to Google maps to further improve its usability. Deep learning algorithms allow the app to extract street names and house numbers from photos. Text recognition in a natural environment is complicated due to distortion, occlusions and other visual artifacts. Machine learning algorithms were challenged on a French Street Name Sign (FSNS) database and can now achieve 84.2% accuracy. When a Street View car drives in a new area, the system analyzes the images collected, extracts street names and numbers. With this information, new addresses are automatically added on Google Maps, continually increasing the accuracy of search results. 

2. Microsoft’s InnerEye for medical diagnosis

Machine learning is being used to help save lives with countless innovations in the healthcare industry. Microsoft is a big player with its InnerEye project aiming to help assist clinicians in treating cancer. InnerEye can help radiologists diagnose and monitor cancer more accurately and efficiently, reducing average treatment time from hours to minutes. The technology analyses radiographic images at the pixel level. This not only tells clinicians where  tumor is but it delineates the organs around surrounding it. Machines use this in radiation therapy where precision is paramount and manual methods previously used were extremely labour intensive.  

3. Timely arrivals from Uber

Machine learning is at the heart of Uber’s business model and is allowing the company to constantly exceed customer expectations. Algorithms are used to accurately predict arrival times, pick-up locations and delivery times. When you book a car, machine learning analyzes data from millions of past trips and calculates the time for your given situation. Data is used from every possible source such as travelling conditions, sincerity level of drivers and time required to prepare different types of food. The end result is that when you book with Uber, it’s estimated time is highly accurate. 

4. Lyst’s fashion-forward search engine

Lyst is a fashion search engine offering millions of fashion products from over 9,000 of the world’s leading brands. With such a huge inventory of products, machine learning is fundamental to the user experience. Machine learning helps users browse Lyst’s products by using metadata tags to make visual comparisons between items of clothing. Due to the high volume of items in the data set, recommendations based on similar users weren’t accurate enough. The machine learning model being used is based on a matrix factorisation algorithm. Each user and item is made up of a sum of representations of its features and recommendations are made on this basis. 

5. Child’s play at Barbie

Machine learning is being used to make our children’s toys come to life. In a reality that seemed an unachievable dream during many of our childhoods, Hello Barbie is an example of a fully interactive toy. A microphone in the toy’s necklace records what the child says and transmits it to the company’s servers. Machine learning is used to choose the correct response from thousands of lines of dialogue. The process is so quick that the response is transmitted back in under a second. 

6. Music compositions from Watson BEAT

Proving that machine learning can be used for much more than data and analytics, IBM’s Watson BEAT has truly ventured into the creative industry. The machine comes up with musical elements to inspire composers and to help them understand what their audiences want. Reinforcement learning creates reward functions by using modern western music theory while a deep belief network is trained to build upon simple input melodies. The software is now open source, offering the potential for user’s songs to be instant hits. 

7. Fraud detection at American Express

Machine learning technologies are invaluable for the detection and prevention of fraudulent payments at American Express. Traditional techniques for fraud detection are based on rules which, for example, flag unusually large withdrawals or cards being used in foreign countries. Machine learning, on the other hand, analyses vasts amounts of transaction data to identify patterns linked to fraudulent behaviour. Payments can be flagged in real time to stop fraudulent activities in their tracks, saving millions in losses. 

8. BBC’s Talking with Machines

Spoken interfaces have become more and more commonplace with the likes of Siri and Alexa at our fingertips. The BBC is trialling a more connected and personalised service that could become more commonplace on its radio channels. Talking with Machines is a radio story that enables two-way conversations via smart speakers. Users get to answer questions and influence the story. The BBC hopes to discover the best types of creative content for speech-based devices with a view to expanding further in the future. 

9. Music to our ears at Spotify

In a similar way to Lyst’s fashionable recommendations, Spotify uses machine learning to make recommendations. The system figures out the type of music you like and dislike to put forward similar suggestions. It does this using collaborative filtering based on users’ behaviour, natural language processing to analyze text and audio models to analyse the raw audio tracks themselves. To make this format even more user-friendly, the recommendations are packaged up in a ‘Discover Weekly’ playlist. The machine learning algorithms hand pick 30 tracks each week based on your user activity and matched to your taste via meta-tags. 

10. MagicBands at Disney

Disney is using machine learning to take its magical park experience to the next level. Every guest to Disney World is given a MagicBand that is fuelled with machine learning technology. The bands communicate with thousands of sensors to optimize and personalize customer experience. The band acts as your keys, cards and tickets. What this means is that data can constantly stream back and forth, recommending the best times for rides, minimizing queues and even allowing your children to meet their favourite characters as if by magic. 

It is without doubt that more and more businesses will adopt machine learning technology in the future. Gartner has predicted that smart machines will enter mainstream adoption by 2021 and as such, machine learning, will be ingrained all the more deeply into our lives. Businesses embracing the technology have the opportunity to gain competitive advantage and to completely redefine the ways in which they work. 

If you’d like to know how we can help with your business then get in touch, we’d love to chat.