FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2103988429> ?p ?o ?g. }

• W2103988429 abstract "Most existing GP regression algorithms assume a single generative model, leading to poor performance when data are nonstationary, i.e. generated from multiple switching processes. Existing methods for GP regression over non-stationary data include clustering and changepoint detection algorithms. However, these methods require significant computation, do not come with provable guarantees on correctness and speed, and most algorithms only work in batch settings. This thesis presents an efficient online GP framework, GP-NBC, that leverages the generalized likelihood ratio test to detect changepoints and learn multiple Gaussian Process models from streaming data. Furthermore, GP-NBC can quickly recognize and reuse previously seen models. The algorithm is shown to be theoretically sample efficient in terms of limiting mistaken predictions. Our empirical results on two real-world datasets and one synthetic dataset show GP-NBC outperforms state of the art methods for nonstationary regression in terms of regression error and computational efficiency. The second part of the thesis introduces a Reinforcement Learning (RL) algorithm, UCRL-GP-CPD, for multi-task Reinforcement Learning when the reward function is nonstationary. First, a novel algorithm UCRL-GP is introduced for stationary reward functions. Then, UCRL-GP is combined with GP-NBC to create UCRL-GP-CPD, which is an algorithm for nonstationary reward functions. Unlike previous work in the literature, UCRL-GP-CPD does not make distributional assumptions about task generation, does not assume changepoint times are known, and does not assume that all tasks have been experienced a priori in a training phase. It is proven that UCRL-GP-CPD is sample efficient in the stationary case, will detect changepoints in the environment with high probability, and is theoretically guaranteed to prevent negative transfer. UCRL-GP-CPD is demonstrated empirically on a variety of simulated and real domains. Thesis Supervisor: Jonathan P. How Title: Richard C. Maclaurin Professor of Aeronautics and Astronautics" @default.
• W2103988429 created "2016-06-24" @default.
• W2103988429 creator A5030570650 @default.
• W2103988429 date "2014-01-01" @default.
• W2103988429 modified "2023-09-23" @default.
• W2103988429 title "Computationally efficient Gaussian Process changepoint detection and regression" @default.
• W2103988429 cites W1483365869 @default.
• W2103988429 cites W1494912895 @default.
• W2103988429 cites W1505937442 @default.
• W2103988429 cites W1514039123 @default.
• W2103988429 cites W1515851193 @default.
• W2103988429 cites W1546511338 @default.
• W2103988429 cites W155097506 @default.
• W2103988429 cites W1563182555 @default.
• W2103988429 cites W1563184156 @default.
• W2103988429 cites W1572161815 @default.
• W2103988429 cites W1576220905 @default.
• W2103988429 cites W1591897158 @default.
• W2103988429 cites W1626155273 @default.
• W2103988429 cites W1660859515 @default.
• W2103988429 cites W1746819321 @default.
• W2103988429 cites W1914756871 @default.
• W2103988429 cites W1945133117 @default.
• W2103988429 cites W1963375715 @default.
• W2103988429 cites W1974889068 @default.
• W2103988429 cites W1981772964 @default.
• W2103988429 cites W1982803779 @default.
• W2103988429 cites W1984615387 @default.
• W2103988429 cites W1987616874 @default.
• W2103988429 cites W2001814969 @default.
• W2103988429 cites W2001825424 @default.
• W2103988429 cites W2003536328 @default.
• W2103988429 cites W2008491003 @default.
• W2103988429 cites W2008674063 @default.
• W2103988429 cites W2014384147 @default.
• W2103988429 cites W2016360498 @default.
• W2103988429 cites W2019324135 @default.
• W2103988429 cites W2020753891 @default.
• W2103988429 cites W2021736234 @default.
• W2103988429 cites W2037539475 @default.
• W2103988429 cites W2042769182 @default.
• W2103988429 cites W2044535354 @default.
• W2103988429 cites W2044987129 @default.
• W2103988429 cites W2052551665 @default.
• W2103988429 cites W2058306002 @default.
• W2103988429 cites W2067603665 @default.
• W2103988429 cites W2073626940 @default.
• W2103988429 cites W2082691056 @default.
• W2103988429 cites W2085330455 @default.
• W2103988429 cites W2089551119 @default.
• W2103988429 cites W2097381042 @default.
• W2103988429 cites W2098949458 @default.
• W2103988429 cites W2099945315 @default.
• W2103988429 cites W2103012681 @default.
• W2103988429 cites W2104533781 @default.
• W2103988429 cites W2107132870 @default.
• W2103988429 cites W2108995755 @default.
• W2103988429 cites W2110292307 @default.
• W2103988429 cites W2111535365 @default.
• W2103988429 cites W2112182839 @default.
• W2103988429 cites W2113442785 @default.
• W2103988429 cites W2115707882 @default.
• W2103988429 cites W2116459397 @default.
• W2103988429 cites W2118686230 @default.
• W2103988429 cites W2119567691 @default.
• W2103988429 cites W2120705090 @default.
• W2103988429 cites W2121863487 @default.
• W2103988429 cites W2123447947 @default.
• W2103988429 cites W2125074935 @default.
• W2103988429 cites W2125710232 @default.
• W2103988429 cites W2125812768 @default.
• W2103988429 cites W2126386737 @default.
• W2103988429 cites W2127498532 @default.
• W2103988429 cites W2129021903 @default.
• W2103988429 cites W2129564505 @default.
• W2103988429 cites W2129670787 @default.
• W2103988429 cites W2130005627 @default.
• W2103988429 cites W2131824593 @default.
• W2103988429 cites W2132908009 @default.
• W2103988429 cites W2133419240 @default.
• W2103988429 cites W2134540127 @default.
• W2103988429 cites W2136816045 @default.
• W2103988429 cites W2138158112 @default.
• W2103988429 cites W2140135625 @default.
• W2103988429 cites W2140484358 @default.
• W2103988429 cites W2146409231 @default.
• W2103988429 cites W2146930087 @default.
• W2103988429 cites W214861312 @default.
• W2103988429 cites W2150535417 @default.
• W2103988429 cites W2152438488 @default.
• W2103988429 cites W2154032554 @default.
• W2103988429 cites W2154185762 @default.
• W2103988429 cites W2155836220 @default.
• W2103988429 cites W2156974606 @default.
• W2103988429 cites W2157577161 @default.
• W2103988429 cites W2157883120 @default.
• W2103988429 cites W2168947765 @default.
• W2103988429 cites W2169743339 @default.
• W2103988429 cites W2172131460 @default.
• W2103988429 cites W2179284380 @default.

• next