Deep learning provides proactive cyber defense

0
25

The increased pace of high-profile threats (e.g. ransomware) is up to doublefigure growth (15.8%). The result is a dangerous path that will most likely lead to sustained losses for organizations that fall victim to a cyber attack without any gains in defensive powers. Indeed, a report on a 2021 data breach by IBM and the Ponemon Institute reveals that the average cost of a data breach is $4.24 million.

In addition to costs, a cyber attack can cause irreparable damage to a company’s brand, stock price and day-to-day operations. According to a recent Deloitte questionnaire, 32% of respondents cited operational disruption as the biggest impact of a cyber incident or breach. Other consequences cited by the companies surveyed include intellectual property theft (22%), a drop in stock price (19%), loss of reputation (17%) and loss of customer confidence (17%).

Given these significant risks, organizations simply cannot afford to accept the status quo of digital asset protection. “If we are ever going to get ahead of our adversaries, the world needs to change the mindset from detection to prevention,” Caspi says. “Organizations need to change the way they implement security and fight hackers.”

Deep learning can make the difference

So far, many cybersecurity experts have seen machine learning as the most innovative approach to protecting digital assets. But deep learning is ideally suited to changing the way we prevent cybersecurity attacks. Any machine learning tool can be understood and theoretically reversed to introduce a bias or vulnerability that will weaken defenses against attack. Bad actors can also use their own machine learning algorithms to pollute a defensive solution with fake data sets.

Fortunately, deep learning overcomes the limitations of machine learning by bypassing the need for highly skilled and experienced data scientists to manually feed a solution data set. Instead, a deep learning model specifically developed for cybersecurity can absorb and process massive amounts of raw data to fully train the system. These neural networks become autonomous, once trained, and do not require constant human intervention. This combination of raw data-based learning and larger data sets means that deep learning is ultimately able to accurately identify much more complex patterns than machine learning, at much faster speeds.

“Deep learning exceeds any denial list, heuristic, or standard machine learning approach,” said Mirel Sehic, vice president general manager for Honeywell Building Technologies (HBT), a multinational company and provider of aerospace, performance materials and safety and productivity technologies. “The time it takes for a deep learning-based approach to detect a specific threat is much faster than all these elements combined.”

Download the full report.

This content is produced by Insights, the custom content arm of MIT Technology Review. It was not written by the editors of MIT Technology Review.