Advances in variational inference

C Zhang, J Bütepage, H Kjellström… - IEEE transactions on …, 2018 - ieeexplore.ieee.org
Many modern unsupervised or semi-supervised machine learning algorithms rely on
Bayesian probabilistic models. These models are usually intractable and thus require …

Learning deep generative models

R Salakhutdinov - Annual Review of Statistics and Its Application, 2015 - annualreviews.org
Building intelligent systems that are capable of extracting high-level representations from
high-dimensional sensory data lies at the core of solving many artificial intelligence–related …

Virtual adversarial training: a regularization method for supervised and semi-supervised learning

T Miyato, S Maeda, M Koyama… - IEEE transactions on …, 2018 - ieeexplore.ieee.org
We propose a new regularization method based on virtual adversarial loss: a new measure
of local smoothness of the conditional label distribution given input. Virtual adversarial loss …

Message passing algorithms for scalable multitarget tracking

F Meyer, T Kropfreiter, JL Williams, R Lau… - Proceedings of the …, 2018 - ieeexplore.ieee.org
Situation-aware technologies enabled by multitarget tracking will lead to new services and
applications in fields such as autonomous driving, indoor localization, robotic networks, and …

Tilted empirical risk minimization

T Li, A Beirami, M Sanjabi, V Smith - arXiv preprint arXiv:2007.01162, 2020 - arxiv.org
Empirical risk minimization (ERM) is typically designed to perform well on the average loss,
which can result in estimators that are sensitive to outliers, generalize poorly, or treat …

[图书][B] Probabilistic graphical models: principles and techniques

D Koller, N Friedman - 2009 - books.google.com
A general framework for constructing and using probabilistic models of complex systems that
would enable a computer to use available information for making decisions. Most tasks …

Deep unfolding: Model-based inspiration of novel deep architectures

JR Hershey, JL Roux, F Weninger - arXiv preprint arXiv:1409.2574, 2014 - arxiv.org
Model-based methods and deep neural networks have both been tremendously successful
paradigms in machine learning. In model-based methods, problem domain knowledge can …

Graphical models, exponential families, and variational inference

MJ Wainwright, MI Jordan - Foundations and Trends® in …, 2008 - nowpublishers.com
The formalism of probabilistic graphical models provides a unifying framework for capturing
complex dependencies among random variables, and building large-scale multivariate …

[图书][B] Variational algorithms for approximate Bayesian inference

MJ Beal - 2003 - search.proquest.com
The Bayesian framework for machine learning allows for the incorporation of prior
knowledge in a coherent way, avoids overfitting problems, and provides a principled basis …

Constructing free-energy approximations and generalized belief propagation algorithms

JS Yedidia, WT Freeman… - IEEE Transactions on …, 2005 - ieeexplore.ieee.org
Important inference problems in statistical physics, computer vision, error-correcting coding
theory, and artificial intelligence can all be reformulated as the computation of marginal …