Bayesian transfer learning between uniformly modelled Bayesian filters

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F67985556%3A_____%2F21%3A00537103" target="_blank" >RIV/67985556:_____/21:00537103 - isvavai.cz</a>

Výsledek na webu

<a href="http://dx.doi.org/10.1007/978-3-030-63193-2_9" target="_blank" >http://dx.doi.org/10.1007/978-3-030-63193-2_9</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-030-63193-2_9" target="_blank" >10.1007/978-3-030-63193-2_9</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Bayesian transfer learning between uniformly modelled Bayesian filters

Popis výsledku v původním jazyce

We investigate sensor network nodes that sequentially infer states with bounded values, and affected by noise that is also bounded. The transfer of knowledge between such nodes is the principal focus of this chapter. A fully Bayesian framework is adopted, in which the source knowledge is represented by a bounded data predictor, the specification of a formal conditioning mechanism between the filtering nodes is avoided, and the optimal knowledge-conditioned target state predictor is designed via optimal Bayesian decision-making (fullynprobabilistic design). We call this framework Bayesian transfer learning, and derive a sequential algorithm for pairs of interacting, bounded filters. To achieve a tractable, finite-dimensional flow, the outputs of the time step, transfer step and data step are locally projected onto parallelotopic supports. An informal variant of the transfer algorithm demonstrates both strongly positive transfer of high-quality (low variance) source knowledge--improving on a former orthotopically supported variant--as well as rejection of low-quality (high variance) source knowledge, which we call robust transfer.

Název v anglickém jazyce

Bayesian transfer learning between uniformly modelled Bayesian filters

Popis výsledku anglicky

We investigate sensor network nodes that sequentially infer states with bounded values, and affected by noise that is also bounded. The transfer of knowledge between such nodes is the principal focus of this chapter. A fully Bayesian framework is adopted, in which the source knowledge is represented by a bounded data predictor, the specification of a formal conditioning mechanism between the filtering nodes is avoided, and the optimal knowledge-conditioned target state predictor is designed via optimal Bayesian decision-making (fullynprobabilistic design). We call this framework Bayesian transfer learning, and derive a sequential algorithm for pairs of interacting, bounded filters. To achieve a tractable, finite-dimensional flow, the outputs of the time step, transfer step and data step are locally projected onto parallelotopic supports. An informal variant of the transfer algorithm demonstrates both strongly positive transfer of high-quality (low variance) source knowledge--improving on a former orthotopically supported variant--as well as rejection of low-quality (high variance) source knowledge, which we call robust transfer.

Druh

C - Kapitola v odborné knize

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/GA18-15970S" target="_blank" >GA18-15970S: Optimální zpracování externí stochastické znalosti vyjádřené pomocí pravděpodobnostních distribucí</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2021

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název knihy nebo sborníku

Informatics in Control, Automation and Robotics : 16th International Conference, ICINCO 2019 Prague, Czech Republic, July 29-31, 2019, Revised Selected Papers

ISBN

978-3-030-63192-5

Počet stran výsledku

18

Strana od-do

151-168

Počet stran knihy

193

Název nakladatele

Springer

Místo vydání

Cham

Kód UT WoS kapitoly

—