Aram-Alexandre Pooladian

[Submitted on 10 Feb 2026] Title:Blind denoising diffusion models and the blessings of dimensionality View a PDF of the paper titled Blind denoising diffusion models and the blessings of dimensionality, by Zahra Kadkhodaie and 3 other authors View PDF HTML (experimental) Abstract:We analyze, theoretically and empirically, the performance of generative diffusion models based on \emph{blind denoisers}, in which the denoiser is not given the noise amplitude in either the training or sampling...

[Submitted on 1 Nov 2024] Title:Wasserstein Flow Matching: Generative modeling over families of distributions View a PDF of the paper titled Wasserstein Flow Matching: Generative modeling over families of distributions, by Doron Haviv and 3 other authors View PDF HTML (experimental) Abstract:Generative modeling typically concerns the transport of a single source distribution to a single target distribution by learning (i.e., regressing onto) simple probability flows.

[Submitted on 5 Dec 2023 ( v1 ), last revised 18 Apr 2025 (this version, v3)] Title:Algorithms for mean-field variational inference via polyhedral optimization in the Wasserstein space View a PDF of the paper titled Algorithms for mean-field variational inference via polyhedral optimization in the Wasserstein space, by Yiheng Jiang and 2 other authors View PDF HTML (experimental) Abstract:We develop a theory of finite-dimensional polyhedral subsets over the Wasserstein space and optimization...