Pith. sign in

REVIEW 2 cited by

PyVertical: A Vertical Federated Learning Framework for Multi-headed SplitNN

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2104.00489 v3 pith:CSBNAV3T submitted 2021-04-01 cs.LG cs.CRcs.DC

classification cs.LGcs.CRcs.DC
keywords dataframeworkacrossneuralfederatedlearningnetworksowners
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

We introduce PyVertical, a framework supporting vertical federated learning using split neural networks. The proposed framework allows a data scientist to train neural networks on data features vertically partitioned across multiple owners while keeping raw data on an owner's device. To link entities shared across different datasets' partitions, we use Private Set Intersection on IDs associated with data points. To demonstrate the validity of the proposed framework, we present the training of a simple dual-headed split neural network for a MNIST classification task, with data samples vertically distributed across two data owners and a data scientist.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Vertical Federated Continual Learning via Evolving Prototype Knowledge

    cs.LG 2025-02 conditional novelty 6.0 of 10

    V-LETO combines evolving class prototypes with Fisher-information-based parameter freezing to handle class and feature incremental learning in vertical federated learning, reporting large gains over baselines.

  2. ProtoGuard-SL: Prototype Consistency Based Backdoor Defense for Vertical Split Learning

    cs.CR 2026-04 conditional novelty 5.0 of 10

    Class-prototype consistency vectors plus class-conditional conformal filtering cut backdoor ASR to ~0.03–0.08 on CIFAR-10, SVHN, and Bank Marketing while preserving main accuracy.

Pith tools