Uber success story has started in 2009. The company was founded by Travis Kalanick and Garret Hobart in San Francisco. Ride sharing industry growth has extended to 37 countries of the world and 128 cities. It all started with the service which was created for the people who wanted to order limousine but did not have enough money for hiring a personal driver. After the owners realized the unique value of this idea, they started working on the spread of it to other target audiences and managed to transform it into the service of ordering a cheap taxi from any point of the city. The startup has encouraged investments of $1,2 billion. The rideshare market basically did not exist at all, and now the rideshare market size grows faster and faster, extending to new cities and getting more and more users and in their face loyal customers.
Uber wants to change a number of things: how we move, how we work, how city planners and car manufacturers think. Uber creates flying cars and is already undermining the achievements of previous decades in the field of labor law. At first glance, the idea itself seems rather straightforward: an application for a smartphone with one click of a button quickly and easily solves the problem of getting around the city. But many in Silicon Valley think: it’s time to forget about Google and Apple, this is the old guard. To understand what the next stage of technologization of the world will be, you need to understand Uber.
Uber is so deeply embedded in the global transport infrastructure that today it’s enough just to press a button to find out how long it takes to get from point A to point B in Paris at any time of day, or how the traffic flows in Kiev.
Facebook changes our ideas about communication, Amazon – about purchases, Uber – about movement.
Systems installed in Uber drones use machine learning for object recognition, motion planning and distributed training for large models on a large number of graphics processors.
How Uber exploits the benefits of machine learning: the company has ensured that its efforts resulted in hundreds of use cases, thousands of working models and millions of forecasts every second. For three years, Uber has significantly expanded the Department of Defense to several hundred developers, researchers and product managers.
Michelangelo Machine Learning Platform
Success is based on Uber’s own MO platform, Michelangelo.
Full working cycle “Michelangelo”:
An important principle that guides the Michelangelo team is to think of machine learning as software development. This means that any MO platform in a company goes through the same repeated and thorough checks that are used in program development.
Look at the MO model, say, as a compiled program library. With this approach, you will, willy-nilly, follow the training configurations using a verified version control system.
Similarly, without having good tracking tools, Uber has repeatedly come across the fact that the model was introduced into the workflow, but it could not be replaced because the training data or configurations were lost.
To make sure that the program works correctly, it is important to conduct quality tests before launching the software and look after it in action. Therefore, before introducing an MO model, Uber always evaluates it on test samples and monitors its work, making sure that it behaves the same way as in offline tests.
The right people working on the right tasks is what matters most when it comes to creating and scaling quality solutions based on MO. And in the face of a shortage of AI specialists, this is becoming increasingly difficult.
There are about 300 thousand specialists in artificial intelligence in the world, while the number of vacancies exceeds several million. And although this is only a hypothesis – good salaries and social packages, as well as an aggressive hiring policy on the part of firms suggest that good AI workers are badly in short supply.
Uber is rapidly acquiring startups as it gears up for a future filled with autonomous ride-hailing vehicles. One day the company announced the purchase of an artificial intelligence startup Geometric Intelligence to provide the core for a new central AI lab being established in Uber’s San Francisco headquarters. Uber AI Labs dedicate itself to researching AI and machine learning, and look for a solution of using object or scenario recognition with smaller sets of data.
The ride-hailing company has been making headlines lately with its intentions of bringing fully autonomous vehicles into the present. Uber currently runs the advanced technologies center, a self-driving car lab in Pittsburgh, and recently acquired self-driving truck company Otto.
“In spite of notable wins with machine learning in recent years, we are still very much in the early innings of machine intelligence,” said Jeff Holden, chief product officer at Uber, in a blog post. “The formation of Uber AI Labs, to be directed by Geometric’s Founding CEO Gary Marcus, represents Uber’s commitment to advancing the state of the art, driven by our vision that moving people and things in the physical world can be radically faster, safer and accessible to all.”
“If you look into the future, there are going to be step-function changes in artificial intelligence that will affect business models and business opportunities,” Holden said. “We very much want to be a part of that.” [Enterprise IoT Inside]
You will receive new articles from Limestone team. It’s all about software development trends for your business success.