Skip to main content
Search

# Partial Optimal Transport

2 related articles

Subscribe to updates
  1. Federated load event recognition stops sharing semantics unconditionally: FedReST edges an architecture-matched FedAvg by 0.00524 Event F1 on REFIT 2014 at 10% labels

    Federated non-intrusive load monitoring lets households collaborate without centralizing raw power traces, yet sparse local labels and household-specific operating patterns make cross-client semantic evidence unevenly reliable. FedReST splits collaboration into two admissions: the server uses reliability-gated partial optimal transport to accept only semantic mass that both sides support, and each receiving client filters the resulting update through function-preserving gradient projection so locally mastered states are not disturbed. On frozen REFIT 2014 rolling folds with 10% visible labels and matched two-epoch local budgets, corrected-core FedReST shows paired differences of +0.00524 in Event F1 and +0.00104 in Window F1 against an architecture-matched Head+BN FedAvg reference.

  2. arXiv

    A new penalty-based optimization framework brings smooth strongly convex regularizers to partial optimal transport, and its accelerated first-order algorithm achieves lower transport cost, higher sparsity, and faster convergence on color transfer, domain adaptation, and point cloud registration

    Targeting the need for sparse transport plans in partial optimal transport (POT), this work proposes a penalty-based reformulation optimization framework that makes smooth strongly convex regularizers such as quadratic or elastic net usable for POT, and builds on it an accelerated first-order algorithm alternating between smooth updates and simple projection steps; on empirical benchmarks in color transfer, domain adaptation, and point cloud registration, the method consistently achieves lower transport cost, higher sparsity, and faster convergence than established baselines.

You’re all caught up