Literature: Mixture Identifiability

Papers on identifiability, estimation, and learning for mixture-like latent decompositions. This area covers classical finite-mixture identifiability, Gaussian mixture recovery, nonparametric and grouped-sample identifiability, latent-class and topic models, conditional-mixture / mixture-of-experts identifiability, and adjacent auxiliary-variable results that clarify when latent structure is recoverable.

Closely related project note: Background on identification and learning of mixture distributions. Candidate papers to ingest are already cited by name in that background note (Teicher 1963, Yakowitz-Spragins 1968, Allman-Matias-Rhodes 2009, Kalai-Moitra-Valiant 2010, Moitra-Valiant 2010, Hsu-Kakade 2013, Vandermeulen-Scott 2019, Ashtiani et al. 2020, and others).

Papers

• Teicher (1963) — Identifiability of Finite Mixtures
• Yakowitz and Spragins (1968) — On the Identifiability of Finite Mixtures
• Chandra (1977) — On the mixtures of probability distributions
• Jiang and Tanner (1999) — Hierarchical Mixtures-of-Experts for Exponential Family Regression Models: Approximation and Maximum Likelihood Estimation
• Holzmann, Munk, and Gneiting (2006) — Identifiability of Finite Mixtures of Elliptical Distributions
• Allman, Matias, and Rhodes (2009) — Identifiability of parameters in latent structure models with many observed variables
• Kasahara and Shimotsu (2009) — Nonparametric Identification of Finite Mixture Models of Dynamic Discrete Choices
• Kalai, Moitra, and Valiant (2010) — Efficiently Learning Mixtures of Two Gaussians
• Moitra and Valiant (2010) — Settling the Polynomial Learnability of Mixtures of Gaussians
• Hsu and Kakade (2013) — Learning mixtures of spherical Gaussians: moment methods and spectral decompositions
• Arora et al. (2013) — A Practical Algorithm for Topic Modeling with Provable Guarantees
• Hardt and Price (2015) — Tight Bounds for Learning a Mixture of Two Gaussians
• Xiang and Yao (2018) — Semiparametric mixtures of nonparametric regressions
• Hyvarinen, Sasaki, and Turner (2019) — Nonlinear ICA Using Auxiliary Variables and Generalized Contrastive Learning
• Vandermeulen and Scott (2019) — An Operator Theoretic Approach to Nonparametric Mixture Models
• Ashtiani et al. (2020) — Near-optimal Sample Complexity Bounds for Robust Learning of Gaussian Mixtures via Compression Schemes