Optimal design of priors constrained by external predictors

Quinn, Anthony; Krn, Miroslav; Guy, Tatiana V.

doi:10.1016/j.ijar.2017.02.001

Cited by 16 publications

(8 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(ii) Below threshold (log 10 r S /r > 1), the absolute performance quantities (TNSE (38), Figure 5a, and AV (39), Figure 5c) coincide for the Bayesian LSU-UOS methods (FPD-BTL and BCM); the DSMF shows better performance.…”

Section: Comparison Of Fpd-btl Bcm and Dsmf For N = 2 Data Channelsmentioning

confidence: 95%

“…as an empirical distribution of a quantity similar to x t , or some unspecified distributional approximation, etc. [38]. In this paper, involving multiple state filtering tasks, we will assume that f S (x S,t |d S (t − 1)) is the output of the source's synchronized time update at t − 1 (3).…”

Section: Fpd-optimal Bayesian Transfer Learning (Fpd-btl)mentioning

confidence: 99%

“…Although there exists an explicit functional minimizer of (13) (see. [38]), our specific purpose here is to instantiate this FPD-optimal solution for UOS-closed filtering in the source and target tasks, as defined in Section 3.1.…”

Section: Fpd-btl Between a Pair Of Lsu-uos Filtering Tasksmentioning

confidence: 99%

“…where ρ and r are positive scalars and 1 (k) is the k-dimensional unit vector. The figures below graph the dependence of TNSE (38), average volume (AV) (39) and average volume ratio (AVR) (40) against the source-to-target ratio of the observation noise variances, i.e. r S /r.…”

Section: Evaluation Criteriamentioning

confidence: 99%

“…In all settings, ρ = 10 −5 , r = 10 −3 (42), t = 1, t = 50, MC = 500 runs. Performances are evaluated via TNSE (38) and AVR (40), in comparison to the isolated target filter. We observe the following: (i) FPD-BTL exhibits positive transfer above threshold; i.e.…”

Section: Fpd-btl With Multiple Source Tasksmentioning

confidence: 99%

See 4 more Smart Citations

Fully probabilistic design for knowledge fusion between Bayesian filters under uniform disturbances

Pavelková¹,

Jirsa²,

Quinn³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

This paper considers the problem of Bayesian transfer learning-based knowledge fusion between linear state-space processes driven by uniform state and observation noise processes. The target task conditions on probabilistic state predictor(s) supplied by the source filtering task(s) to improve its own state estimate. A joint model of the target and source(s) is not required and is not elicited. The resulting decision-making problem for choosing the optimal conditional target filtering distribution under incomplete modelling is solved via fully probabilistic design (FPD), i.e. via appropriate minimization of Kullback-Leibler divergence (KLD). The resulting FPD-optimal target learner is robust, in the sense that it can reject poor-quality source knowledge. In addition, the fact that this Bayesian transfer learning (BTL) scheme does not depend on a model of interaction between the source and target tasks ensures robustness to the misspecification of such a model. The latter is a problem that affects conventional transfer learning methods. The properties of the proposed BTL scheme are demonstrated via extensive simulations, and in comparison with two contemporary alternatives.

show abstract

Section: Comparison Of Fpd-btl Bcm and Dsmf For N = 2 Data Channelsmentioning

confidence: 95%

Section: Fpd-optimal Bayesian Transfer Learning (Fpd-btl)mentioning

confidence: 99%

Section: Fpd-btl Between a Pair Of Lsu-uos Filtering Tasksmentioning

confidence: 99%

Section: Evaluation Criteriamentioning

confidence: 99%

Section: Fpd-btl With Multiple Source Tasksmentioning

confidence: 99%

See 3 more Smart Citations