IBM Watson receives federated learning and time series updates

Get a free Techzine subscription!

IBM is making updates to the Watson AI technology, for business use. The company will add support for federated learning and time series capabilities in Watson Studio. The changes are designed to help customers build AI solutions they can trust, according to Big Blue.

The company also said that the new capabilities expand on Watson tools designed to be used by businesses in governing and explaining AI-led decisions, increase the accuracy of insights offered and mitigate risks while meeting the privacy and compliance requirements.

IBM’s data science platform, Watson Studio, now comes with a federated learning capability update.

What the updates do

The federated learning capability is only still a technical preview but, it will be used to train AI models on data sources that would have previously been in separate silos. Such silos may be where data cannot or should not be moved because of reasons like compliance, privacy, secrecy, or even volume.

Watson Studio also has beta support for time series capabilities, to solve problems like analysis, automation, and forecasting time series data, typically seen in industries like retail, finance, and manufacturing.

IBM added that Time Series on Watson Studio is designed to help achieve accuracy on datasets like travel time, call data logs, retail sales, and more.

What else is new?

IBM OpenPages with Watson now comes with a Data Privacy Management module, which can be used to meet the changing data privacy problems.

Integrating the Watson Knowledge Catalog into it can provide businesses with an almost real-time view of how private data is being utilized by their organizations.

IBM Planning Analytics with Watson will gain a new statistical details page for transparency and digestible facts about how forecasting was generated. The features should be available soon, with the Analytics update arriving sometime in the second quarter of this year.