Establish patient-rights features in FL-Net

Establish informed consent, automated revocation notification, and model-based unlearning as first-class platform features of FL-Net in order to address patient rights in the federated clinical research infrastructure.

Background

FL-Net currently addresses patient rights indirectly through cohort-based permissions, data-discovery-query thresholds, local patient traceability, and auditable change logs. The paper identifies informed consent, automated notification when consent is revoked, and model-based unlearning as distinct capabilities that are not yet integrated as first-class platform functions.

Resolving this problem would extend FL-Net beyond technical data minimization toward stronger normative and operational support for data-protection rights, including consent management, withdrawal of consent, and removal of a patient’s influence from trained models.

References

Informed consent, automated revocation notification, and model-based unlearning, on the other hand, have not yet been established as first-class Platform features.

— Multi-center Medical Data Mining with FL-Net - A One-stop Shop for Federated Learning  (2609.20650 - Süwer et al., 17 Sep 2026) in Section “Auditing” (Methods)

As a problem-space signal, federated unlearning is the open question of how to guarantee that a client's contribution can be removed from a trained federated model and how to verify that the removal took place.

— BackTrend: Evaluating Scientific Weak-Signal Prediction via Backward Reconstruction  (2609.24921 - Zhou et al., 21 Sep 2026) in Appendix, Section “Two Cases That Warrant Explanation,” paragraph “A single direction can occupy both signal spaces of one topic”