AWS Entity Resolution introduces incremental machine learning-based matching workflows

AWS Entity Resolution has officially launched its support for incremental machine learning (ML) based matching workflows, now available for general use. This advancement marks a significant shift in how businesses can handle entity resolution on a large scale. Traditionally, the addition of even a single new record necessitated the reprocessing of an entire dataset, a task that could take up to two days and incur substantial costs. This inefficiency has been a major obstacle for enterprises, often leading them to explore expensive alternatives or workarounds.

The newly introduced feature by AWS Entity Resolution allows organizations to focus exclusively on processing new records that have been added since their last workflow execution. This change brings about remarkable improvements in efficiency, with the capability to process one million incremental records in under an hour. This represents a 95% reduction in processing time compared to existing workflows, while also significantly lowering infrastructure expenses.

The incremental matching feature can accommodate workloads of up to 50 million new records, alongside datasets containing up to one billion historical base records. This makes AWS Entity Resolution a feasible option for ongoing, large-scale enterprise tasks that were previously considered too costly to implement.

The incremental ML workflows are now accessible in all AWS Regions where AWS Entity Resolution is available. For those interested in initiating an incremental ML workflow, detailed instructions can be found in the user guide. Additional information about AWS Entity Resolution can be accessed on the product page.