Quantum Fourier generative models are trained classically at over 1000-qubit scale using log-likelihood loss from Parseval's identity and deployed on superconducting hardware for fast sampling that preserves multi-modal structure.
Generalized Denoising Auto-Encoders as Generative Models
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
abstract
Recent work has shown how denoising and contractive autoencoders implicitly capture the structure of the data-generating density, in the case where the corruption noise is Gaussian, the reconstruction error is the squared error, and the data is continuous-valued. This has led to various proposals for sampling from this implicitly learned density function, using Langevin and Metropolis-Hastings MCMC. However, it remained unclear how to connect the training procedure of regularized auto-encoders to the implicit estimation of the underlying data-generating distribution when the data are discrete, or using other forms of corruption process and reconstruction errors. Another issue is the mathematical justification which is only valid in the limit of small corruption noise. We propose here a different attack on the problem, which deals with all these issues: arbitrary (but noisy enough) corruption, arbitrary reconstruction loss (seen as a log-likelihood), handling both discrete and continuous-valued variables, and removing the bias due to non-infinitesimal corruption noise (or non-infinitesimal contractive penalty).
years
2026 3verdicts
UNVERDICTED 3representative citing papers
Extends variable projection to constrained separable nonlinear least-squares via bilevel collapse, yielding exact reduced gradients and a convergent conditional-gradient algorithm.
Separating presence from magnitude in sparse temporal audit data lets a dual-channel autoencoder focus learning on anomalous activity for insider threat detection.
citing papers explorer
-
Quantum Fourier Generative Models Trainable at Large Scale
Quantum Fourier generative models are trained classically at over 1000-qubit scale using log-likelihood loss from Parseval's identity and deployed on superconducting hardware for fast sampling that preserves multi-modal structure.
-
Constrained Variable Projection for Structured Problems
Extends variable projection to constrained separable nonlinear least-squares via bilevel collapse, yielding exact reduced gradients and a convergent conditional-gradient algorithm.
-
Signal Decomposition Reveals Structure in Insider Threat Detection under Sparse Temporal Data
Separating presence from magnitude in sparse temporal audit data lets a dual-channel autoencoder focus learning on anomalous activity for insider threat detection.