Matches in SemOpenAlex for { <https://semopenalex.org/work/W3135077735> ?p ?o ?g. }
- W3135077735 abstract "Across machine learning, the use of curricula has shown strong empirical potential to improve learning from data by avoiding local optima of training objectives. For reinforcement learning (RL), curricula are especially interesting, as the underlying optimization has a strong tendency to get stuck in local optima due to the exploration-exploitation trade-off. Recently, a number of approaches for an automatic generation of curricula for RL have been shown to increase performance while requiring less expert knowledge compared to manually designed curricula. However, these approaches are seldomly investigated from a theoretical perspective, preventing a deeper understanding of their mechanics. In this paper, we present an approach for automated curriculum generation in RL with a clear theoretical underpinning. More precisely, we formalize the well-known self-paced learning paradigm as inducing a distribution over training tasks, which trades off between task complexity and the objective to match a desired task distribution. Experiments show that training on this induced distribution helps to avoid poor local optima across RL algorithms in different tasks with uninformative rewards and challenging exploration requirements." @default.
- W3135077735 created "2021-03-15" @default.
- W3135077735 creator A5017983137 @default.
- W3135077735 creator A5063752250 @default.
- W3135077735 creator A5066979765 @default.
- W3135077735 creator A5071367253 @default.
- W3135077735 creator A5080021372 @default.
- W3135077735 creator A5088706700 @default.
- W3135077735 date "2021-02-25" @default.
- W3135077735 modified "2023-09-29" @default.
- W3135077735 title "A Probabilistic Interpretation of Self-Paced Learning with Applications to Reinforcement Learning" @default.
- W3135077735 cites W130216483 @default.
- W3135077735 cites W1499669280 @default.
- W3135077735 cites W1523436756 @default.
- W3135077735 cites W158722652 @default.
- W3135077735 cites W1771410628 @default.
- W3135077735 cites W181733065 @default.
- W3135077735 cites W1863227302 @default.
- W3135077735 cites W1983628095 @default.
- W3135077735 cites W1991247894 @default.
- W3135077735 cites W1995137594 @default.
- W3135077735 cites W1996847178 @default.
- W3135077735 cites W2012587148 @default.
- W3135077735 cites W2013164703 @default.
- W3135077735 cites W2024060531 @default.
- W3135077735 cites W2045724293 @default.
- W3135077735 cites W2080039641 @default.
- W3135077735 cites W2097381042 @default.
- W3135077735 cites W2098774185 @default.
- W3135077735 cites W2100913937 @default.
- W3135077735 cites W2101524054 @default.
- W3135077735 cites W2107042471 @default.
- W3135077735 cites W2121275167 @default.
- W3135077735 cites W2123157758 @default.
- W3135077735 cites W2132984949 @default.
- W3135077735 cites W2138537392 @default.
- W3135077735 cites W2140801763 @default.
- W3135077735 cites W2145339207 @default.
- W3135077735 cites W2152166054 @default.
- W3135077735 cites W2158782408 @default.
- W3135077735 cites W2167117957 @default.
- W3135077735 cites W2170156068 @default.
- W3135077735 cites W2171578145 @default.
- W3135077735 cites W2208154600 @default.
- W3135077735 cites W2211399972 @default.
- W3135077735 cites W2213381658 @default.
- W3135077735 cites W2256388387 @default.
- W3135077735 cites W2296073425 @default.
- W3135077735 cites W2417786368 @default.
- W3135077735 cites W2419612459 @default.
- W3135077735 cites W2551887912 @default.
- W3135077735 cites W2561776174 @default.
- W3135077735 cites W2620239873 @default.
- W3135077735 cites W2724149175 @default.
- W3135077735 cites W2733961795 @default.
- W3135077735 cites W2736601468 @default.
- W3135077735 cites W2737215407 @default.
- W3135077735 cites W2751302235 @default.
- W3135077735 cites W2766447205 @default.
- W3135077735 cites W2785738552 @default.
- W3135077735 cites W2788191306 @default.
- W3135077735 cites W2793798239 @default.
- W3135077735 cites W2799151646 @default.
- W3135077735 cites W2902334345 @default.
- W3135077735 cites W2914275007 @default.
- W3135077735 cites W2945024121 @default.
- W3135077735 cites W2949608212 @default.
- W3135077735 cites W2949900957 @default.
- W3135077735 cites W2951747194 @default.
- W3135077735 cites W2954700257 @default.
- W3135077735 cites W2963293881 @default.
- W3135077735 cites W2964118020 @default.
- W3135077735 cites W2964118044 @default.
- W3135077735 cites W2964161785 @default.
- W3135077735 cites W2970599228 @default.
- W3135077735 cites W2997289589 @default.
- W3135077735 cites W3028935481 @default.
- W3135077735 cites W3032597593 @default.
- W3135077735 cites W3037745193 @default.
- W3135077735 cites W3037760466 @default.
- W3135077735 cites W3085438811 @default.
- W3135077735 cites W3099671035 @default.
- W3135077735 cites W3123298421 @default.
- W3135077735 cites W3145164839 @default.
- W3135077735 cites W3208165232 @default.
- W3135077735 cites W43241812 @default.
- W3135077735 cites W567721252 @default.
- W3135077735 hasPublicationYear "2021" @default.
- W3135077735 type Work @default.
- W3135077735 sameAs 3135077735 @default.
- W3135077735 citedByCount "0" @default.
- W3135077735 crossrefType "posted-content" @default.
- W3135077735 hasAuthorship W3135077735A5017983137 @default.
- W3135077735 hasAuthorship W3135077735A5063752250 @default.
- W3135077735 hasAuthorship W3135077735A5066979765 @default.
- W3135077735 hasAuthorship W3135077735A5071367253 @default.
- W3135077735 hasAuthorship W3135077735A5080021372 @default.
- W3135077735 hasAuthorship W3135077735A5088706700 @default.
- W3135077735 hasConcept C119857082 @default.
- W3135077735 hasConcept C12713177 @default.