Fed-Universe proposes negotiated alignment for federated learners
Original title: Fed-Universe: A Semantic-Geometric-Topological-Human (S-G-T-H) Stack for Negotiated Alignment in Federated Systems
A synthetic binational clinical simulation cut a minority node's loss from 0.857 to 0.340 despite under 3% of the data.
Fed-Universe proposes a semantic, geometric, topological and human stack so federated learning negotiates alignment instead of averaging toward a global mean. An edge LLM turns heterogeneous inputs into patient-profile prompts, and later layers gate minority updates and balance objectives. Authors report a synthetic binational clinical simulation in which a node with under 3% of the data reduced loss from 0.857 to 0.340. Expansion into Fed-Ultra and Fed-Human is described as ongoing.
Unconfirmed: The reported parity comes from a synthetic simulation, not a real clinical deployment.
- The minority node held less than 3% of data volume.
- Loss on that node fell from 0.857 to 0.340.
- Authors say the framework is expanding into Fed-Ultra and Fed-Human.
- Products
- Fed-UniverseFed-UltraFed-Human
From the original article
1 min read at arxiv.orgConventional federated learners implicitly optimize for global means, creating a "one-size-fits-all" paradigm that inevitably suppresses the minority under cross-site heterogeneity. To bridge this critical gap, we present Fed-Universe, a generalizable Semantic-Geometric-Topological-Human (S-G-T-H) architecture that transforms passive averaging into an active negotiation for decentralized alignment.
By Ting Xu · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.