Better Programming Saved Netflix $1 Billion in 2019 While only a small portion of Netflix's development budget goes towards its recommendation engine, the returns have been astronomical. If Uribe-Gomez and Hunt's assessment that the recommendation engine saves the company more than $1 billion per year is even close to accurate, it's seeing a significant return on investment from its AI research. This is the power of Machine Learning Because of newer computing technologies, machine learning today is not like machine learning of the 70s.
It was born of pattern recognition and the theory that computers can study without being programmed to do specific tasks; researchers working on artificial intelligence wanted to see if computers could study from data. The iterative viewpoint of machine learning is important: as models are exposed to new data, they are able to autonomously adapt. They learn from earlier computations to produce strong, repeatable decisions and results. It’s a science that’s not new – but one that has obtained recent momentum. Same-day shipping from Amazon is available because of machine learning. In fact, their current ML algorithm has decreased the ‘click-to-ship’ time by 225%. The newest systems also make it easier to learn how machines can learn. If classical programming turns rules and data into answers, machine learning algorithms will work backward and turn answers and data into rules. They can give you insights about what is going on in the depths of your business. The developers of these simplified tools are also building interfaces that illustrate the rules that the algorithm discovered and, more importantly, how to duplicate the results. The power of the tools is in their ability to handle the grungy work of collecting data, appending structure and consistency where possible, and then begin the calculation. They clarify the data gathering process and the labor of keeping the information in rows and columns. But tools are still not very smart. They just can’t do all of this learning for you. You might ask the right questions, look in the right places and insightfully analyze results that were produced. All the machines can do is speed up the search process and help you cut some corners.
By data investigation and discovery, ML software can quickly extract insights from data and shows them in understandable(“Humane”) formats. The software can lower a bar of knowledge, so even beginners can do something meaningful and useful. To expand your ML knowledge or to seek help in rethinking your ML workflow, check outComet.ml (this newsletter’s sponsor) and start tracking your ML experiments. You can start experimenting and use tools to allow people to build ML models together. But the quality of the ML software should be high. Sometimes you can’t rely on some open-sourced library - especially if you need to use your results at production. The clearness, qualitative code, collaboration options, performance, and general documentation is written in plain language provide to project success and good analyze. An additional plus for ML software is an ability to work with the research task and iterative machine learning tasks. The interface is important as well. And if you can separate some of the tasks - it’s also good for ML researchers - it helps to apply changes more quickly. Sometimes quick drag-n-drop works better than quick code adjustment...
In the meanwhile, check out Hacker Noon curated stories from machine learning professionals.
If you have a few (or many) experiences to share, you can become a contributor on Hacker Noon too. Simply create an account, and join 10,000+ contributors sharing their knowledge and expertise with the rest of us. Maybe you'll get featured on our next newsletter too.
