Fujitsu, Linux Foundation Partner on Open-Source A.I.

(Seanbatty, Pixabay)

15 Sept. 2023. I.T. systems company Fujitsu Ltd. is providing artificial intelligence model generation and fairness assessment software as open-source projects hosted by the Linux Foundation. The Linux Foundation in San Francisco says it approved incubating the two initiatives in late August, and announced the collaboration ahead of the organization’s Open Source Summit Europe conference that begins next week in Bilbao, Spain.

Fujitsu, based in Tokyo, offers software from its A.I. platform code-named Fujitsu Kozuchi, unveiled in April 2023.  The company says Fujitsu Kozuchi is designed for collaborative and interactive development with customers that generates core models such as machine learning and generative A.I., with additional business components applied as needed. The added business components offer features such as defect inspection and analysis, fraud monitoring, and consumer analytics. Fujitsu says it holds 970 patents related to A.I.

Find an optimum solution and predictive model

Linux Foundation and Fujitsu seek to encourage further rapid development of A.I., but acknowledge the need to make A.I. tools more accessible to users without advanced I.T. skills, as well as ensuring all user populations are accurately reflected in algorithms. The SapientML project, one of the joint initiatives, generates machine learning models and code from data presented in tabular form. Data sets in tables train the models to produce a workflow for predicting outcomes in a new data collection. SapientML works in successive stages to identify and test machine learning components, winnowing them down to construct a process for finding the optimum solution and a predictive model. The software was presented in May 2022 at the International Conference on Software Engineering.

Intersectional Fairness, the other Fujitsu project, seeks to detect and mitigate biases against population groups in A.I. algorithms that may be hidden in model training data. The process, as described in an earlier paper, assesses the interaction of population attributes in outcomes such as gender or age with comparative statistical tests and real-world data sets to identify biases in training data. The model further evaluates the combination of those attributes and adjusts for biases without impairing accuracy of the underlying model.

Both the SapientML and Intersectional Fairness projects are available as open-source software or OSS on GitHub. Fujitsu Kozuchi includes automated machine learning model generation and ethics-for-fairness among its core modules. “Offering A.I. technologies as OSS to developers worldwide,” says Fujitu’s chief technology officer Vivek Mahajan in a Linux Foundation statement, “opens up new opportunities for innovation across various industries by lowering the barrier of entry.” Mahajan is scheduled to give one of the keynote speeches at the Open Source Summit Europe conference next week.

More from Science & Enterprise:

• A.I. Analytics Detect Heart Failure in Single Ultrasound View
• Venture Fund to Back Life Science A.I. Start-Ups
• Infographic – U.S. Public Backs A.I. Regulation
• Automated Life Science A.I. Generation System Unveiled
• Companies Partner on A.I., Fusion Energy for Space Flight

We designed Science & Enterprise for busy readers including investors, researchers, entrepreneurs, and students. Except for a narrow cookies and privacy strip for first-time visitors, we have no pop-ups blocking the entire page, nor distracting animated GIF graphics. If you want to subscribe for daily email alerts, you can do that here, or find the link in the upper left-hand corner of the desktop page. The site is free, with no paywall. But, of course, donations are gratefully accepted.

*     *     *