Since the concept of “machines learning” was introduced in the 1950s, the field has gone from a cryptic domain understood by a few (Turing, Markov, Legendre, Laplace or Bayes) to a technology that every company must deploy.
Every day we hear how data and automation improve our shopping experiences, our online searches and enables fraud prevention and cybersecurity routines to do more, faster and better for us.
Now, the amalgamates created around Artificial Intelligence, Machine Learning and Big Data are bound to confuse industry observers or investors who aren’t familiar with the technical details. If you’re asking yourself: “What’s the difference between Big Data and Machine Learning?”, then for the sake of my piece, simply think about it this way: “Big Data is Machine Learning’s great uncle”. Machine Learning doesn’t need Big Data to exist. But, if it uses it, it can benefit greatly from its vast knowledge.
Big Data is Machine Learning’s great uncle. Machine Learning doesn’t need Big Data to exist. But, if it uses it, it can benefit greatly from its vast knowledge"
I’m sure you’ll find more sophisticated answers out there but what matters most is not just the technology. It’s how technology is applied. So rather than discuss typical enterprise use cases that might bore you to death, I thought I’d highlight interesting (and somewhat unexpected) scenarios where data and machine learning play a role.
Drones as flying data collectors
You might not think of drones as having anything to do with data and machine learning. But they’re beginning to have a huge impact in a number of industries from mining to construction to farming.
“We see ourselves as a data company, not a drone company,” says Airobotics CEO Ran Krauss. For Airobotics, which just announced the opening of a US headquarters in Arizona, the biggest market today for its fully automated drone system is the mining industry. Its drones are becoming increasingly common around active mining areas, providing accurate 3D models and 2D maps on a daily basis.
"We supply autonomous drones that facilitate a range of services, depending on the needs of our clients, from security to surveying. Whatever the drone is doing, it’s constantly collecting data from the different points of interest within the mine site. This data supports both routine and critical, time-sensitive mining operations, while providing important business insights, introducing cost savings, and improving safety,” says Krauss.
The mining industry has already begun to look at all sorts of uses for Big Data. Big Data may soon be utilized to both aid the discovery of new mineral deposits and provide safer mining operations. For example, by using real-time monitoring of sensors on miners, their environment, and equipment, hyper-fast analytics may identify risks such as an impending tunnel collapse, or deteriorating air quality within the mine.
3D Printing and Big Data poised to reinvent manufacturing
In just the last couple of years, the popularity of 3D printing has exploded, but it’s been mostly as a source of made-to-order items rather than mass production parts. 3D printing has also made rapid prototyping of parts for the aerospace, automotive, and medical industries possible.
The challenge for mass production is quality control and that’s where researchers say data and machine learning will come in. Some recent research on the use of sensor-based, in-process monitoring has been done, with the goal of giving 3D printing machines the ability to recognize malicious defects in real-time.
The key here is machine learning. Big Data was collected during the building process so that the system could learn to spot anomalies in real-time and provide automated self-correction during the manufacturing process.
“Unlike 3D printing for private use, mass manufacturing demands a very high level of quality, and more importantly, consistency of quality,” says Avi Reichental, founder of XponentialWorks, and former CEO of 3D Systems. “If we’re talking about components for the auto or aircraft industries, for example, it’s not enough that they look good; they must be flawless, every time.”
Will car companies become data companies?
The automotive industry has been using data and process automation for some time in many aspects of car manufacturing, parts distribution, and sales. There’s nothing new about that. But, a former director of Israel’s security service surprised attendees of the Paris Motor Show last month when he suggested that in the near future car companies will become Big Data companies.
"The car industry is undergoing a dramatic revolution,” Yuval Diskin, the former director of the Israeli Internal Security Service (Shin Bet) and a co-founder of cybersecurity startup CyMotive Technologies Ltd, told Calcalist Tech. “Now they are starting to understand that IT is the core business, because with time they will have more and more data, and the automotive industry will become a Big Data industry.”
What sort of data? Hundreds of internal sensors are already collecting data about your car. In the near-future, self-driving cars will also generate hundreds of terabytes of real-time data about road and traffic conditions, traffic volumes, and even weather reports.
All of that Data is going to put demands on storage and analysis, especially in areas like predictive analytics and anomaly detection. The payoff will be much safer cars and an all round better driving experience, with the cars themselves providing the data for better roads and more efficient cars. For that reason, as reported in Forbes, auto manufacturers are already looking for ways to monetize that data.
The days of Big Data hype are over. We’re now at the stage of real enterprise adoption across so many verticals, you might be surprised at the ways Big Data is affecting you without your even realizing it. This trend is only going to continue as more exciting and disruptive applications are applied in every industry sector.
