Deep Learning in Data Protection, Cybersecurity, and Ransomware Prevention

We often reference Arcserve Unified Data Protection (UDP) and its protection and prevention capabilities across on- and off-premises workloads in these posts. We also usually note that Arcserve UDP is “safeguarded by Sophos Intercept X Advanced cybersecurity, uniquely combining deep-learning server protection, immutable storage, and scalable onsite and offsite business continuity.”

While we’ve illustrated the importance of immutable storage and scalability, we’ve never explained what deep learning—one of the key drivers behind Intercept X Advanced for Server’s cybersecurity capabilities—is and why it matters. As Sophos writes on its website, deep learning makes Intercept X for Server smarter, more scalable, and more effective against never-before-seen threats, adding that deep learning outperforms security solutions that use traditional machine learning or signature-based detection alone.

Deep Learning vs. Machine Learning

It’s important to understand that deep learning is a subset of machine learning. The key differences are in how each algorithm learns and how much data each uses. Deep learning eliminates some manual human intervention by automating most of the feature extraction steps of the process. It can also use large data sets and has been referred to as “scalable machine learning.”

Gartner estimates that unstructured data adds up to as much as 80 to 90 percent of all new enterprise data, growing at a rate three times faster than structured data. That’s important because typical machine learning requires more structured data to learn, depending on human intervention—human experts—to determine the hierarchy of features that help delineate the differences between data inputs and guide understanding of the data. These experts decide which characteristics matter for each dataset, using labels or distinguishing features to streamline the learning process through supervised learning.

Deep learning—in reality, deep machine learning—can leverage supervised learning but doesn’t require it. Instead, deep learning can ingest unstructured data in its raw form (text or images, for example) and automatically determine the set of features that distinguish each dataset. By observing patterns in data, a deep learning model can correctly cluster inputs, grouping datasets based on the similarities or differences it has identified.

Still, a deep learning model may require more data points to improve its accuracy, while machine learning relies on less data. That’s why deep learning is mainly employed for complex use cases—like cybersecurity. That’s why Arcserve partnered with Sophos to ensure we include the most effective prevention measures with our products.

Smarter Ransomware Defenses

Arcserve relies on Intercept X for Server for its anti-ransomware capabilities that detect malicious encryption processes and block them before they can spread across your network—including file-based and master boot record (MBR) ransomware attacks.

Cloud-native security offers advanced protection and visibility across your cloud, on-premises, and virtual server workloads. And Intercept X Advanced delivers extended detection and response, going beyond the server by incorporating cross-product data sources for increased visibility. The software blocks exploits and techniques used to distribute malware, steal credentials, and escape detection.

Unfortunately, between human error and social engineering, hackers can still find a way to get past your defenses. That’s where Arcserve UDP has you covered, offering protection against data loss and extended downtime across cloud, local, virtual, hyperconverged, and SaaS-based workloads using a single unified management interface. It reduces downtime from days to minutes and validates your RPOs, RTOs, and service-level agreements (SLAs) with automated testing and granular reporting.

Discover all the benefits of Arcserve UDP with integrated Sophos cybersecurity software by checking out our on-demand demo and 30-day free trial.