【转】机器学习在制造业的应用场景

原文地址：http://www.mbtmag.com/article/2016/06/enhancing-manufacturing-operations-artificial-intelligence

A few different scenarios whereby deep learning can interpret mass quantities of information to provide “signals” to a manufacturer that point to potential areas of improvement.

Customer monitoring: In manufacturing, hundreds of millions of dollars’ worth of orders can go out the door with payment hinging on a customer’s ability to pay. Machine learning can provide the real-time monitoring manufacturers need to sift through news articles, blogs and other unstructured content to determine if a client is experiencing negative financial performance or other triggering indicators that may signal an inability to pay.

 

Feedback loop: By using artificial intelligence to track the performance of a specific product or application in the marketplace, manufacturers can spot problems before they become a customer relations nightmare. Ongoing scanning and analysis of structured and unstructured information can piece together key indicators of a manufacturing defect or other anomaly that can be remedied before it becomes a widespread problem.

 

Lead generation: Using the same logic as above, artificial intelligence can also do the deep digging needed to spot signals for future product needs that would allow a manufacturer to proactively re-tool or otherwise prepare for incoming orders. This level of ongoing news and data analysis would not be productive for a human worker, but use of AI coupled with contextual machine learning can provide the steady, ongoing consumption of information needed to spot new business opportunities buried deep in stacks of information.

 

There is tremendous interest and need for timely analysis and interpretation of the massive amount of unstructured data [along with structured data] that is being created in the new digital age. By adopting an approach that utilizes fine-grained linguistic based analysis, inaccuracies and coarse-grained interpretation that is common among conventional platforms can be avoided. By focusing on all types of structured and unstructured data, enterprises can leverage rapid analysis of untold volumes of information for business advantage without tying up valuable knowledge workers in the process.

转载于:https://www.cnblogs.com/victoria693/p/6800188.html