analiza aiqbench
This paper advances blind source separation by proposing MuDo-CoM, an algorithm capable of identifying dependent latent components across multiple domains using only second-order statistics. By relaxing the independence assumptions standard in Independent Component Analysis (ICA), this method offers improved robustness for complex real-world scenarios like multi-sensor fusion and fMRI analysis where underlying factors are inherently correlated.
Identifying Dependent Components from Multi-domain Linear Mixtures
arXiv:2610.28330v1 Announce Type: new Abstract: We study a linear mixing model with dependent latent components, assuming multiple data domains. Most existing models assume that the components are independent or at least uncorrelated, in line with independent component analysis (ICA). Some recent work allows for dependent components, but then makes specific assumptions such as parametric forms of dependencies, multi-view settings, or interventions, or does not recover the individual components. In contrast, we consider a multi-domain setting in which domains differ through domain-specific scalings of the components, while the distribution of the underlying latent components is the same across domains. This approach can model data collected, for example, from different sensors measuring the same process, different laboratories conducting the same experiment, different experimental conditions, or different subjects that might differ in biological or physiological factors. We show that, under sufficient domain variability, latent variables and mixing functions can be identified from second-order statistics alone. We propose the Multi-Domain Covariance Matching (MuDo-CoM) algorithm that generalizes previous methods of joint diagonalization. MuDo-CoM is validated on simulated data and a real-world fMRI dataset.
Abstract:We study a linear mixing model with dependent latent components, assuming multiple data domains. Most existing models assume that the components are independent or at least uncorrelated, in line with independent component analysis (ICA). Some recent work allows for dependent components, but then makes specific assumptions such as parametric forms of dependencies, multi-view settings, or interventions, or does not recover the individual components. In contrast, we consider a multi-domain setting in which domains differ through domain-specific scalings of the components, while the distribution of the underlying latent components is the same across domains. This approach can model data collected, for example, from different sensors measuring the same process, different laboratories conducting the same experiment, different experimental conditions, or different subjects that might differ in biological or physiological factors. We show that, under sufficient domain variability, latent variables and mixing functions can be identified from second-order statistics alone. We propose the Multi-Domain Covariance Matching (MuDo-CoM) algorithm that generalizes previous methods of joint diagonalization. MuDo-CoM is validated on simulated data and a real-world fMRI dataset.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2610.28330 [cs.LG]
(or arXiv:2610.28330v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2610.28330 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 270 cuvinte
Subiecte conexe
Unsupervised Multidomain Approaches to Named Entity Recognition with Small Datasets
arXiv · 4 aug. 2026
Beyond AHI: An Interpretable Causal-Discovery-Guided Framework for Sleep Recovery in Connected Health
arXiv · 2 sept. 2026
FAIRChem v2 UMA for Multidomain Atomistic Simulation across Molecules, Catalysts, Materials, Vibrations, and Molecular Dynamics
MarkTechPost · 26 iul. 2026