FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3204256165> ?p ?o ?g. }

• W3204256165 abstract "Biological constraints often impose restrictions for plausible plasticity rules such as locality and reward-based rather than supervised learning. Two learning rules that comply with these restrictions are weight (WP) and node (NP) perturbation. NP is often used in learning studies, in particular as a benchmark; it is considered to be superior to WP and more likely neurobiologically realized, as the number of weights and therefore their perturbation dimension typically massively exceeds the number of nodes. Here we show that this conclusion no longer holds when we take two biologically relevant properties into account: First, tasks extend in time. This increases the perturbation dimension of NP but not WP. Second, tasks are low dimensional, with many weight configurations providing solutions. We analytically delineate regimes where these properties let WP perform as well as or better than NP. Furthermore we find that the changes in weight space directions that are irrelevant for the task differ qualitatively between WP and NP and that only in WP gathering batches of subtasks in a trial decreases the number of trials required. This may allow to experimentally distinguish which of the two rules underlies a learning process. Our insights suggest new learning rules, which combine for specific task types the advantages of WP and NP. If the inputs are similarly correlated, temporally correlated perturbations improve NP. Using numerical simulations, we generalize the results to networks with various architectures solving biologically relevant and standard network learning tasks. Our findings, together with WP’s practicability suggest WP as a useful benchmark and plausible model for learning in the brain." @default.
• W3204256165 created "2021-10-11" @default.
• W3204256165 creator A5013951895 @default.
• W3204256165 creator A5035501683 @default.
• W3204256165 creator A5069942719 @default.
• W3204256165 date "2021-10-05" @default.
• W3204256165 modified "2023-10-06" @default.
• W3204256165 title "Weight Perturbation Learning Performs Similarly or Better than Node Perturbation on Broad Classes of Temporally Extended Tasks" @default.
• W3204256165 cites W1927368590 @default.
• W3204256165 cites W1965742370 @default.
• W3204256165 cites W1970585834 @default.
• W3204256165 cites W1970890515 @default.
• W3204256165 cites W1984465702 @default.
• W3204256165 cites W1995367941 @default.
• W3204256165 cites W1996739924 @default.
• W3204256165 cites W2008008116 @default.
• W3204256165 cites W2014295524 @default.
• W3204256165 cites W2017002109 @default.
• W3204256165 cites W2017539895 @default.
• W3204256165 cites W2019982479 @default.
• W3204256165 cites W2024529445 @default.
• W3204256165 cites W2025687881 @default.
• W3204256165 cites W2028526571 @default.
• W3204256165 cites W2029053566 @default.
• W3204256165 cites W2031096147 @default.
• W3204256165 cites W2045529778 @default.
• W3204256165 cites W2050579567 @default.
• W3204256165 cites W2052210110 @default.
• W3204256165 cites W2061897041 @default.
• W3204256165 cites W2062670775 @default.
• W3204256165 cites W2063046089 @default.
• W3204256165 cites W2065196330 @default.
• W3204256165 cites W2066835186 @default.
• W3204256165 cites W2071527746 @default.
• W3204256165 cites W2084794931 @default.
• W3204256165 cites W2088947792 @default.
• W3204256165 cites W2095031087 @default.
• W3204256165 cites W2095312056 @default.
• W3204256165 cites W2095708593 @default.
• W3204256165 cites W2099231026 @default.
• W3204256165 cites W2101295242 @default.
• W3204256165 cites W2106308939 @default.
• W3204256165 cites W2106437793 @default.
• W3204256165 cites W2109302401 @default.
• W3204256165 cites W2116424792 @default.
• W3204256165 cites W2116524281 @default.
• W3204256165 cites W2119717200 @default.
• W3204256165 cites W2124001356 @default.
• W3204256165 cites W2124290836 @default.
• W3204256165 cites W2132753014 @default.
• W3204256165 cites W2136287399 @default.
• W3204256165 cites W2145255740 @default.
• W3204256165 cites W2156084351 @default.
• W3204256165 cites W2162160137 @default.
• W3204256165 cites W2162407586 @default.
• W3204256165 cites W2162933103 @default.
• W3204256165 cites W2165971697 @default.
• W3204256165 cites W2168648327 @default.
• W3204256165 cites W2171111726 @default.
• W3204256165 cites W2171658832 @default.
• W3204256165 cites W2172115975 @default.
• W3204256165 cites W2183261201 @default.
• W3204256165 cites W2321192406 @default.
• W3204256165 cites W2497568564 @default.
• W3204256165 cites W2552737632 @default.
• W3204256165 cites W2606461235 @default.
• W3204256165 cites W2613472289 @default.
• W3204256165 cites W2613867659 @default.
• W3204256165 cites W2621961142 @default.
• W3204256165 cites W2726666509 @default.
• W3204256165 cites W2753961618 @default.
• W3204256165 cites W2766284800 @default.
• W3204256165 cites W2780064148 @default.
• W3204256165 cites W2914670473 @default.
• W3204256165 cites W2943879006 @default.
• W3204256165 cites W2945612185 @default.
• W3204256165 cites W2950141843 @default.
• W3204256165 cites W2950800702 @default.
• W3204256165 cites W2953293408 @default.
• W3204256165 cites W2954137290 @default.
• W3204256165 cites W3002831475 @default.
• W3204256165 cites W3013369344 @default.
• W3204256165 cites W3016391357 @default.
• W3204256165 cites W3081530306 @default.
• W3204256165 cites W3133994644 @default.
• W3204256165 cites W3162006056 @default.
• W3204256165 cites W386569623 @default.
• W3204256165 cites W4212952892 @default.
• W3204256165 cites W4231081240 @default.
• W3204256165 cites W4242770316 @default.
• W3204256165 doi "https://doi.org/10.1101/2021.10.04.463055" @default.
• W3204256165 hasPublicationYear "2021" @default.
• W3204256165 type Work @default.
• W3204256165 sameAs 3204256165 @default.
• W3204256165 citedByCount "0" @default.
• W3204256165 crossrefType "posted-content" @default.
• W3204256165 hasAuthorship W3204256165A5013951895 @default.
• W3204256165 hasAuthorship W3204256165A5035501683 @default.
• W3204256165 hasAuthorship W3204256165A5069942719 @default.
• W3204256165 hasBestOaLocation W32042561651 @default.

• next