Federated learning lets multiple parties train a shared model without pooling raw data. The new paper introduces NOMAI, a scalable federated learning framework that the authors say is designed to protect the parties contributing data.

According to the abstract, NOMAI pursues three properties at once: model privacy, data privacy, and integrity. The authors frame these as the twin goals of privacy-preserving federated learning — harnessing distributed data while protecting contributors. The abstract is truncated, so the specific mechanism is not described in the available text.

With only a single source and an incomplete abstract, this summary is limited to the stated goals. No experimental results or technical comparisons are reported in the source.