Matches in SemOpenAlex for { <https://semopenalex.org/work/W3010106586> ?p ?o ?g. }
- W3010106586 abstract "Latent variable models represent observed variables as parameterized functions of a set of latent variables. Examples are factor analysis or probabilistic sparse coding which assume weighted linear summations to determine the mean of Gaussian distribution for the observables. However, in many cases observables do not follow a normal distribution, and a linear summation of latents is often at odds with non-Gaussian observables (e.g., means of the Bernoulli distribution have to lie in the unit interval). Furthermore, the assumption of a linear summation model may (for many types of data) not be closely aligned with the true data generation process even for Gaussian observables. Alternative superposition models (i.e., alternative links between latents and observables) have therefore been investigated repeatedly. Here we show that using the maximization instead of summation to link latents to observables allows for the derivation of a very general and concise set of parameter update equations. Concretely, we derive a set of update equations that has the same functional form for all distributions of the exponential family. Our results consequently provide directly applicable learning equations for commonly as well as for unusually distributed data. We numerically verify our analytical results assuming standard Gaussian, Gamma, Poisson, Bernoulli and Exponential distributions. We point to some potential applications by providing different experiments on the learning of variance structure, noise type estimation, and denoising." @default.
- W3010106586 created "2020-03-13" @default.
- W3010106586 creator A5003594160 @default.
- W3010106586 creator A5022342273 @default.
- W3010106586 creator A5031051637 @default.
- W3010106586 creator A5060530231 @default.
- W3010106586 date "2020-03-04" @default.
- W3010106586 modified "2023-10-14" @default.
- W3010106586 title "Maximal Causes for Exponential Family Observables." @default.
- W3010106586 cites W1490256991 @default.
- W3010106586 cites W1510608507 @default.
- W3010106586 cites W1512208174 @default.
- W3010106586 cites W1520448186 @default.
- W3010106586 cites W1545527235 @default.
- W3010106586 cites W1562453744 @default.
- W3010106586 cites W1732796048 @default.
- W3010106586 cites W1761606332 @default.
- W3010106586 cites W1902027874 @default.
- W3010106586 cites W1906770428 @default.
- W3010106586 cites W1964155876 @default.
- W3010106586 cites W2012762214 @default.
- W3010106586 cites W2021262032 @default.
- W3010106586 cites W2034926629 @default.
- W3010106586 cites W2037642501 @default.
- W3010106586 cites W2041465237 @default.
- W3010106586 cites W2044784704 @default.
- W3010106586 cites W2048695508 @default.
- W3010106586 cites W2056370875 @default.
- W3010106586 cites W2066973160 @default.
- W3010106586 cites W2071005004 @default.
- W3010106586 cites W2083380015 @default.
- W3010106586 cites W2098062695 @default.
- W3010106586 cites W2098567366 @default.
- W3010106586 cites W2101933716 @default.
- W3010106586 cites W2103633133 @default.
- W3010106586 cites W2106884367 @default.
- W3010106586 cites W2113046264 @default.
- W3010106586 cites W2115938734 @default.
- W3010106586 cites W2117824861 @default.
- W3010106586 cites W2120340025 @default.
- W3010106586 cites W2125027820 @default.
- W3010106586 cites W2128209899 @default.
- W3010106586 cites W2128514094 @default.
- W3010106586 cites W2135001774 @default.
- W3010106586 cites W2135046866 @default.
- W3010106586 cites W2138102231 @default.
- W3010106586 cites W2138311613 @default.
- W3010106586 cites W2141522908 @default.
- W3010106586 cites W2145889472 @default.
- W3010106586 cites W2148495331 @default.
- W3010106586 cites W2160547390 @default.
- W3010106586 cites W2161133254 @default.
- W3010106586 cites W2161164955 @default.
- W3010106586 cites W2162427980 @default.
- W3010106586 cites W2162929365 @default.
- W3010106586 cites W2216511711 @default.
- W3010106586 cites W2257718385 @default.
- W3010106586 cites W2296616510 @default.
- W3010106586 cites W2404032589 @default.
- W3010106586 cites W2421886322 @default.
- W3010106586 cites W2508457857 @default.
- W3010106586 cites W2530211015 @default.
- W3010106586 cites W2567948266 @default.
- W3010106586 cites W2576174010 @default.
- W3010106586 cites W2576674590 @default.
- W3010106586 cites W2582082042 @default.
- W3010106586 cites W2657268502 @default.
- W3010106586 cites W2740438624 @default.
- W3010106586 cites W2757860842 @default.
- W3010106586 cites W2770581584 @default.
- W3010106586 cites W2810198369 @default.
- W3010106586 cites W2891092832 @default.
- W3010106586 cites W2907700557 @default.
- W3010106586 cites W2910720383 @default.
- W3010106586 cites W2951122352 @default.
- W3010106586 cites W2962885243 @default.
- W3010106586 cites W2963368172 @default.
- W3010106586 cites W2963704386 @default.
- W3010106586 cites W2964125708 @default.
- W3010106586 cites W2964193753 @default.
- W3010106586 cites W2997123853 @default.
- W3010106586 cites W3023343579 @default.
- W3010106586 cites W3119963126 @default.
- W3010106586 cites W3140968660 @default.
- W3010106586 cites W44831635 @default.
- W3010106586 cites W629338568 @default.
- W3010106586 cites W755141289 @default.
- W3010106586 hasPublicationYear "2020" @default.
- W3010106586 type Work @default.
- W3010106586 sameAs 3010106586 @default.
- W3010106586 citedByCount "0" @default.
- W3010106586 crossrefType "posted-content" @default.
- W3010106586 hasAuthorship W3010106586A5003594160 @default.
- W3010106586 hasAuthorship W3010106586A5022342273 @default.
- W3010106586 hasAuthorship W3010106586A5031051637 @default.
- W3010106586 hasAuthorship W3010106586A5060530231 @default.
- W3010106586 hasConcept C105795698 @default.
- W3010106586 hasConcept C121332964 @default.
- W3010106586 hasConcept C121864883 @default.
- W3010106586 hasConcept C122123141 @default.