FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3207538905> ?p ?o ?g. }

• W3207538905 abstract "Bayesian Likelihood-Free Inference (LFI) approaches allow to obtain posterior distributions for stochastic models with intractable likelihood, by relying on model simulations. In Approximate Bayesian Computation (ABC), a popular LFI method, summary statistics are used to reduce data dimensionality. ABC algorithms adaptively tailor simulations to the observation in order to sample from an approximate posterior, whose form depends on the chosen statistics. In this work, we introduce a new way to learn ABC statistics: we first generate parameter-simulation pairs from the model independently on the observation; then, we use Score Matching to train a neural conditional exponential family to approximate the likelihood. The exponential family is the largest class of distributions with fixed-size sufficient statistics; thus, we use them in ABC, which is intuitively appealing and has state-of-the-art performance. In parallel, we insert our likelihood approximation in an MCMC for doubly intractable distributions to draw posterior samples. We can repeat that for any number of observations with no additional model simulations, with performance comparable to related approaches. We validate our methods on toy models with known likelihood and a large-dimensional time-series model." @default.
• W3207538905 created "2021-10-25" @default.
• W3207538905 creator A5018689327 @default.
• W3207538905 creator A5060445865 @default.
• W3207538905 date "2020-12-20" @default.
• W3207538905 modified "2023-09-27" @default.
• W3207538905 title "Score Matched Neural Exponential Families for Likelihood-Free Inference" @default.
• W3207538905 cites W112870199 @default.
• W3207538905 cites W1505878979 @default.
• W3207538905 cites W1522301498 @default.
• W3207538905 cites W1546037769 @default.
• W3207538905 cites W1565709818 @default.
• W3207538905 cites W1587348082 @default.
• W3207538905 cites W1680396847 @default.
• W3207538905 cites W1793259860 @default.
• W3207538905 cites W1836465849 @default.
• W3207538905 cites W1959736062 @default.
• W3207538905 cites W1964607942 @default.
• W3207538905 cites W1970391837 @default.
• W3207538905 cites W1978556363 @default.
• W3207538905 cites W1979685007 @default.
• W3207538905 cites W1979969656 @default.
• W3207538905 cites W1984048068 @default.
• W3207538905 cites W1994129134 @default.
• W3207538905 cites W2013035813 @default.
• W3207538905 cites W2025341678 @default.
• W3207538905 cites W2025720061 @default.
• W3207538905 cites W2030911724 @default.
• W3207538905 cites W2034956278 @default.
• W3207538905 cites W2043983054 @default.
• W3207538905 cites W2045973738 @default.
• W3207538905 cites W2051144468 @default.
• W3207538905 cites W2051588269 @default.
• W3207538905 cites W2052329135 @default.
• W3207538905 cites W2062125287 @default.
• W3207538905 cites W2092124742 @default.
• W3207538905 cites W2116064496 @default.
• W3207538905 cites W2119135250 @default.
• W3207538905 cites W2121794292 @default.
• W3207538905 cites W2123368367 @default.
• W3207538905 cites W2135267747 @default.
• W3207538905 cites W2138204974 @default.
• W3207538905 cites W2139812092 @default.
• W3207538905 cites W2152246075 @default.
• W3207538905 cites W2167030304 @default.
• W3207538905 cites W2212660284 @default.
• W3207538905 cites W2343952582 @default.
• W3207538905 cites W2534708821 @default.
• W3207538905 cites W2577537660 @default.
• W3207538905 cites W2581158580 @default.
• W3207538905 cites W2593094133 @default.
• W3207538905 cites W2626930952 @default.
• W3207538905 cites W2753890080 @default.
• W3207538905 cites W2763783893 @default.
• W3207538905 cites W2768551975 @default.
• W3207538905 cites W2774549550 @default.
• W3207538905 cites W2804609220 @default.
• W3207538905 cites W2903442788 @default.
• W3207538905 cites W2908411392 @default.
• W3207538905 cites W2914562565 @default.
• W3207538905 cites W2948142124 @default.
• W3207538905 cites W2949236913 @default.
• W3207538905 cites W2951779732 @default.
• W3207538905 cites W2963295938 @default.
• W3207538905 cites W2963297889 @default.
• W3207538905 cites W2963399751 @default.
• W3207538905 cites W2963822196 @default.
• W3207538905 cites W2963918832 @default.
• W3207538905 cites W2963956018 @default.
• W3207538905 cites W2963993553 @default.
• W3207538905 cites W2964224666 @default.
• W3207538905 cites W2964250089 @default.
• W3207538905 cites W2964273226 @default.
• W3207538905 cites W2964378242 @default.
• W3207538905 cites W2970803184 @default.
• W3207538905 cites W2970971581 @default.
• W3207538905 cites W2975024646 @default.
• W3207538905 cites W2978753741 @default.
• W3207538905 cites W2992035660 @default.
• W3207538905 cites W3000092734 @default.
• W3207538905 cites W3002551964 @default.
• W3207538905 cites W3004228280 @default.
• W3207538905 cites W3007285736 @default.
• W3207538905 cites W3010959735 @default.
• W3207538905 cites W3012496101 @default.
• W3207538905 cites W3014325088 @default.
• W3207538905 cites W3034706502 @default.
• W3207538905 cites W3036815251 @default.
• W3207538905 cites W3037981576 @default.
• W3207538905 cites W3038048211 @default.
• W3207538905 cites W3038056456 @default.
• W3207538905 cites W3086988147 @default.
• W3207538905 cites W3091100449 @default.
• W3207538905 cites W3093431131 @default.
• W3207538905 cites W3094221767 @default.
• W3207538905 cites W3104594340 @default.
• W3207538905 cites W3105637242 @default.
• W3207538905 cites W3106513449 @default.
• W3207538905 hasPublicationYear "2020" @default.
• W3207538905 type Work @default.

• next