Machine learning building blocks create complexity, says Wikibon analyst

man machine brain gears magnifying glass ai artificial intelligence thought

Advanced analysis tools are critical to extracting value from big data. The problem, writes Wikibon Big Data & Analytics Analyst George Gilbert, is that the various open source and proprietary building blocks of machine learning analysis applications often use incompatible application and administrative services, making it difficult to build smooth-running applications.

This complexity is slowing adoption of big data deployments in all but the largest, most technically sophisticated organizations. For example, trusted systems require authentication, authorization and data governance. Each of these is developed independently, either as open source projects or by private vendors. Each development team takes its own approach to the problem, exposes its own, often proprietary application program interfaces and generally does not consider compatibility with the other building blocks.

The same is true for the other three management categories: scalability, resilience and automation. These are far from plug-and-play, and the result is that machine learning applications work inefficiently at best, and too often break along the boundaries between incompatible elements.

In the full Professional Alert, available to Wikibon Premium subscribers, Gilbert examines the problems in each category of service. His recommendation is to partner with a cloud provider when building and deploying big data applications.

To find out about becoming a Wikibon subscriber, look here. Wikibon is a sister company to SiliconANGLE.

Image by “geralt” via Pixabay