Machine Learning, like the other popular buzzwords these days, is in for a major growth spurt. The market is expected to cross $30 million by 2025 as per latest research. And while most sectors want to get on the bandwagon, a challenge they still face is identifying a use case – which is feasible, cost effective and yields returns.
If someone posts a photo on Twitter, they want it to be seen. But if the thumbnail is not cropped right, there is a chance no one will. Twitter has been using neural networks to resolve this problem. Its algorithm uses machine learning to learn how to crop photos into low-resolution, but compelling previews. The result is a scalable, cost-effective solution which shows lesser number of tires and more interesting number plates!
Apple filed a patent some time back for cross-device personalization. Club that with the amazing new health features in the watch series 4 – where it can track your heartbeat better, and even record an ECG. In the future you might be seeing your device suggesting a playlist according to your current activity level! Working out, you get a playlist with upbeat music. Relaxing, might want to listen to something more soothing!
Target had a very famous “incident” where they suggested baby products to a teenager based on their machine learning algorithms. Agreed, that was a bit too much but it does tell you how well machine learning algorithms can predict user purchase behavior. Machine learning can pick up the best time to pitch certain products to customers based on their recent purchases – and not just based on the season or upcoming holidays. Consider a new car owner – they might be a good target to sell car accessories to. By using machine learning effectively with Big Data, retailers can recommend products to customers when they are most likely to buy it.
A large number of people purchase products online using retail giants like Alibaba and Amazon. And their recommendation engine is on-point. But how do they track so many user journeys – their searches and purchases – with machine learning. The home pages on their websites are customized for shoppers based on their past purchases, search results show the products people are looking for. Alibaba even has a chatbot which handles customer queries. And these algorithms keep learning – every action a user takes gives it more inputs to base it future results on.
While it is true that machines can’t truly replace humans or human interaction. But companies like Apple, Twitter etc are pushing the boundaries. With machine learning and big data making innovation easier than ever, we are excited to see how companies use them.