M+E Connections

Google Cloud Touts Growing Importance of Machine Learning

NEW YORK — The increasing importance of machine learning (ML) across many industries was stressed at the Google Cloud OnBoard New York Big Data and Machine Learning Edition training event Feb. 21.

“No matter which industry you’re working [in] or what you do as your daily job, machine learning is no longer a strange term to us anymore,” Han Wang, customer engineer and machine learning specialist at Google, said during a presentation early at the event.

“So, what is actually machine learning?” she asked rhetorically, adding: “To explain in layman’s terms, machine learning is a way that we build a system [and] the system will learn by itself over time. Machine learning is a branch of artificial intelligence. It is actually a way that we solve problems where we don’t have to explicitly codify the solution.”

Wang went on to show examples of ML use cases in different industries and how they leverage the technology.

Among the many industries now using ML, she noted, are: media (where it’s used for content recommendation, searching and auto news writing); manufacturing of many devices (for quality assurance, process optimization, demand forecasting and product innovation); retail (for customer relationship improvement and shopping experience personalization); healthcare (to diagnose diseases from real-time patient data, outcome prediction, acute detection, precision medicine and device management); education and training (for teaching assistants, performance prediction and evaluation, and customized learning); and financial services (for fraud detection, contact centers, searching, file indexing and credit references).

Citing McKinsey & Company data, she said applied ML for quality assurance can help improve semiconductor manufacturing yields up to 30%, supply chain forecasting errors to be reduced 50%, inventory costs to be reduced 20-50%, and — best of all — to increase defect detection rates by a whopping 90%, significantly improving the automation of quality testing.

“These numbers show us great opportunity and benefits of applying machine learning to our industry,” she said. She singled out the significant way in which ML has helped internet searches, which we all do every day, noting that when we performed a search of something in the past, what we received was “exactly what you searched” for. Now, however, when you type in a word such as “Giants” using Google, you’re going to see “some relative content” that will suggest specific things you may be looking for, including San Francisco Giants, where to buy tickets and what time that day’s baseball game starts, she noted.
What that indicates is that the development of ML technology is happening “faster than ever,” she said.

She went on to predict that, “in the next five years, machine learning in each industry will be increasing at a dramatic speed.”

Citing a prediction from fellow Media & Entertainment Services Alliance (MESA) member PwC, she said manufacturers’ adoption of ML and analytics to improve predictive maintenance is expected to improve from 28% today to 66% in the next five years.

Based on what she’s seen, she predicted “a lot more enterprises and organizations are going to adopt machine learning” in the years to come.