Wasserstein
PulseAugur coverage of Wasserstein — every cluster mentioning Wasserstein across labs, papers, and developer communities, ranked by signal.
7 day(s) with sentiment data
-
New stable filters enhance generative models for graph signals
Researchers have developed a new framework for designing stable graph filters to improve generative models for graph signals. These filters are designed to preserve the smoothing properties of graph heat diffusion while…
-
New research explores faster convergence in AI sampling methods · 2 sources tracked
Researchers have published new findings on Wasserstein-Fisher-Rao (WFR) gradient flows, a method for accelerating convergence in sampling from probability distributions. The latest work, building on previous research, a…
-
New math paper details nonlocal transport convergence rates
A new paper published on arXiv details a mathematical framework for understanding nonlocal transport phenomena. The research focuses on the convergence of solutions to a nonlocal continuity equation towards heat flow, e…
-
New geometry framework enhances machine learning metrics
Researchers have developed a new framework for generalized infinite-dimensional Alpha-Procrustes based geometries, extending existing metrics like Bures-Wasserstein and Log-Euclidean. This formalism, based on unitized H…
-
New ParetoTransport method enhances generative optimization for multi-objective problems
Researchers have introduced ParetoTransport, a novel training-free guidance method for generative models aimed at improving offline multi-objective optimization. This method explicitly refines the distribution of candid…
-
New DRIFT attack removes diffusion model watermarks with 98-100% success
Researchers have developed a new method called DRIFT to remove watermarks embedded in images generated by diffusion models. This black-box attack works by partially re-noising images and then using stochastic reverse re…
-
D-FROST algorithm uses optimal transport for decentralized prompt tuning
Researchers have introduced D-FROST, a novel decentralized federated learning algorithm designed for prompt tuning. This method addresses challenges in decentralized settings, such as non-aligned prompt sets and the nee…
-
New study quantifies generative augmentation reliability using Wasserstein discrepancy
A new study published on arXiv explores the theoretical impact of generative data augmentation on downstream generalization in machine learning. The research introduces a statistical framework to analyze how augmentatio…
-
New theory quantifies convergence for Langevin-regularized SVGD
This paper introduces a new theoretical framework for understanding Langevin-regularized Stein Variational Gradient Descent (SVGD). The research establishes quantitative convergence guarantees to the target distribution…
-
New method infers artist relationships from music acoustics
Researchers have developed a method to infer expert critic-sourced network adjacency between musical artists by analyzing acoustic distributions. This approach uses optimal-transport distances on acoustic descriptors to…
-
New `findr` framework balances credit risk accuracy with fairness and transparency
Researchers have introduced `findr`, a novel semi-structured framework designed for binary credit risk modeling. This framework aims to balance predictive accuracy with transparency and fairness by decomposing the logit…
-
New deep learning framework enhances causal inference for multi-treatment scenarios
Researchers have developed CIHSI-Net, a deep learning framework designed to improve causal inference for heterogeneous treatment effects under multiple simultaneous treatments. The framework utilizes a novel Barycentric…
-
New MS-WDRO framework fuses heterogeneous graph data using Wasserstein metric
Researchers have developed a novel framework called MS-WDRO for learning graph structures from multiple, heterogeneous data sources. This method leverages the Wasserstein metric to fuse diverse datasets by calculating a…
-
New risk quantification methods enhance AI agent safety under uncertainty
Researchers have developed new methods for agents to quantify and manage risk in uncertain environments, particularly during the learning phase. One approach, RATTL (Risk-Adversarial Total-Reward Learning), ties caution…
-
New AI framework BRaG learns stock trading from diverse expert strategies
Researchers have developed BRaG, a novel framework for stock trading that utilizes adversarial inverse reinforcement learning to learn from diverse expert strategies. This approach aggregates heterogeneous trading style…
-
New EMS Coreset algorithm offers efficient data subsetting for machine learning
Researchers have developed EMS Coreset, a novel algorithm designed to create representative data subsets for machine learning tasks more efficiently. This method utilizes an expectation-maximization approach with Sinkho…
-
New algorithm enhances stability for mean-field variational inference
Researchers have developed a finite-batch particle algorithm for mean-field variational inference, extending its stability beyond strongly convex potentials. The algorithm is analyzed as a discrete approximation of proj…
-
New research explores robust distribution learning with Wasserstein metrics · 3 sources tracked
Three new research papers explore advanced methods for robust distribution learning and online optimization using Wasserstein metrics. The first paper introduces Wasserstein Filtering (WF) to identify and remove contami…
-
New Wasserstein Policy Gradient Method for LQ Control Problems
Researchers have developed a Wasserstein policy gradient (WPG) method for entropy-regularized linear-quadratic (LQ) control problems. This approach leverages the fact that unrestricted LQ control problems have linear-Ga…
-
New sub-quadratic method improves bisimulation metric computation for MDPs
Researchers have developed a novel sub-quadratic method for calculating bisimulation metrics in Markov decision processes (MDPs). This new approach utilizes approximate nearest neighbor (ANN) indexing to efficiently sel…