Wasserstein metric
PulseAugur coverage of Wasserstein metric — every cluster mentioning Wasserstein metric across labs, papers, and developer communities, ranked by signal.
3 天有情绪数据
-
机器学习模型因数据偏移导致能见度预测困难
研究人员开发了一个机器学习框架，用于预测韩国六个城市的空气能见度，解决了数据不平衡和分布偏移等挑战。该研究采用了SMOTENC和CTGAN等技术来处理数据不平衡，并使用机器学习和深度学习模型的集成进行预测。与交叉验证相比，测试集上的性能显著下降，突显了时间分布偏移的影响，该影响使用Wasserstein距离进行了量化。
-
New method optimizes Wasserstein distance estimation runtime
Researchers have developed a new method to optimize the computational-statistical runtime for estimating Wasserstein distance. This technique, called Sample-Sketch-Solve, uses a regular cartesian grid to sketch data, wh…
-
New research links Föllmer processes to DDPMs, improving sampling efficiency
Researchers have explored the connection between Föllmer processes and denoising diffusion probabilistic models (DDPMs), finding that discretizing Föllmer processes can yield optimal hyper-parameter settings for DDPM sa…
-
New GANICE method advances causal inference with Wasserstein distance
Researchers have introduced GANICE, a new method for distributional causal inference that utilizes Generative Adversarial Networks (GANs) to estimate interventional outcome distributions. This approach addresses limitat…
-
New DP sampling method uses Wasserstein distance
Researchers have introduced a new framework for differentially private sampling from distributions, utilizing Wasserstein distance as the primary utility measure. This approach addresses limitations of prior methods tha…
-
CoMemNet improves continual traffic prediction with memory replay and contrastive sampling
Researchers have introduced CoMemNet, a novel dual-branch continual learning framework designed for traffic prediction in dynamic, evolving networks. This system employs an Online branch for immediate predictions and a …
-
New Fisher Decorator method refines offline RL policies with local transport maps
Researchers have developed a new method called Fisher Decorator to improve flow-based offline reinforcement learning. This approach addresses limitations in existing methods by using a local transport map to refine poli…
-
New research analyzes full-graph vs. mini-batch GNN training
This paper presents a comprehensive analysis comparing full-graph and mini-batch training for Graph Neural Networks (GNNs). It explores the impact of batch size and fan-out size on GNN convergence and generalization, of…
-
New methods enhance conformal prediction for uncertainty quantification
Researchers have developed novel methods for conformal prediction, a technique used for uncertainty quantification in machine learning. The first approach utilizes a differentiable nonconformity score to create a flow o…
-
Deep Neural Networks Achieve Universality via Lindeberg Exchange Principle
Researchers have developed a new approach to understand the behavior of deep neural networks in their infinite-width limit. By applying a Lindeberg principle specifically adapted for deep neural networks, they can quant…
-
New Bayes Posterior Sampling Method Enhances Large-Data Mixed Models
Researchers have developed a novel stochastic mirror Langevin dynamics algorithm designed for fitting Bayesian generalized linear mixed models to large datasets. This new method addresses limitations in existing stochas…
-
New method uses hidden states to improve AI reasoning credit assignment
Researchers have developed a new method called Span-level Hidden state Enabled Advantage Reweighting (SHEAR) to improve credit assignment in reinforcement learning for language models. SHEAR leverages the Wasserstein di…
-
New Bayesian design framework improves experimental efficiency using integral probability metrics
Researchers have developed a new Bayesian Optimal Experimental Design (BOED) framework that utilizes integral probability metrics (IPMs) to enhance stability and accuracy. This approach replaces traditional Kullback-Lei…