RECORD
Preconditioning the prior to overcome saturation in Bayesian inverse problems
Statistics Theory · Submitted 23 Sep 2014
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim We show how to overcome this saturation in an empirical Bayesian framework by using a non-centered data-dependent prior.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
We study Bayesian inference in statistical linear inverse problems with Gaussian noise and priors in Hilbert space. We focus our interest on the posterior contraction rate in the small noise limit. Existing results suffer from a certain saturation phenomenon, when the data generating element is too smooth compared to the smoothness inherent in the prior. We show how to overcome this saturation in an empirical Bayesian framework by using a non-centered data-dependent prior. The center is obtained from a preconditioning regularization step, which provides us with additional information to be used in the Bayesian framework. We use general techniques known from regularization theory. To highlight the significance of the findings we provide several examples. In particular, our approach allows to obtain and, using preconditioning improve after saturation, minimax rates of contraction established in previous studies. We also establish minimax contraction rates in cases which have not been considered so far.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | We study Bayesian inference in statistical linear inverse problems with Gaussian noise and priors in Hilbert space. | background 0.87 | 0.50 | 0.02 | asserted 0.82 |
| s1 | We focus our interest on the posterior contraction rate in the small noise limit. | background 0.70 | 0.45 | 0.03 | asserted 0.71 |
| s2 | Existing results suffer from a certain saturation phenomenon, when the data generating element is too smooth compared to the smoothness inherent in the prior. | gap 0.64 | 0.04 | 0.04 | asserted 0.99 |
| s3 | We show how to overcome this saturation in an empirical Bayesian framework by using a non-centered data-dependent prior. | contribution 0.76 | 0.95 | 0.76 | asserted 1.00 |
| s4 | The center is obtained from a preconditioning regularization step, which provides us with additional information to be used in the Bayesian framework. | method 0.99 | 0.77 | 0.09 | asserted 1.00 |
| s5 | We use general techniques known from regularization theory. | method 0.90 | 0.32 | 0.02 | asserted 0.86 |
| s6 | To highlight the significance of the findings we provide several examples. | contribution 0.47 | 0.84 | 0.03 | asserted 0.92 |
| s7 | In particular, our approach allows to obtain and, using preconditioning improve after saturation, minimax rates of contraction established in previous studies. | result 0.86 | 0.85 | 0.37 | asserted 0.86 |
| s8 | We also establish minimax contraction rates in cases which have not been considered so far. | result 0.73 | 0.95 | 0.17 | asserted 1.00 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
We study Bayesian inference in statistical linear inverse problems with Gaussian noise and priors in Hilbert space. We focus our interest on the posterior contraction rate in the small noise limit. Existing results suffer from a certain saturation phenomenon, when the data generating element is too smooth compared to the smoothness inherent in the prior. We show how to overcome this saturation in an empirical Bayesian framework by using a non-centered data-dependent prior. The center is obtained from a preconditioning regularization step, which provides us with additional information to be used in the Bayesian framework. We use general techniques known from regularization theory. To highlight the significance of the findings we provide several examples. In particular, our approach allows to obtain and, using preconditioning improve after saturation, minimax rates of contraction established in previous studies. We also establish minimax contraction rates in cases which have not been considered so far.
Human review
How the literature uses it
2 arXiv papers cite this in 4 sentences
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.