Matches in SemOpenAlex for { <https://semopenalex.org/work/W3196052105> ?p ?o ?g. }
- W3196052105 abstract "Exploration remains a central challenge for reinforcement learning (RL). Virtually all existing methods share the feature of a monolithic behaviour policy that changes only gradually (at best). In contrast, the exploratory behaviours of animals and humans exhibit a rich diversity, namely including forms of switching between modes. This paper presents an initial study of mode-switching, non-monolithic exploration for RL. We investigate different modes to switch between, at what timescales it makes sense to switch, and what signals make for good switching triggers. We also propose practical algorithmic components that make the switching mechanism adaptive and robust, which enables flexibility without an accompanying hyper-parameter-tuning burden. Finally, we report a promising and detailed analysis on Atari, using two-mode exploration and switching at sub-episodic time-scales." @default.
- W3196052105 created "2021-08-30" @default.
- W3196052105 creator A5016618355 @default.
- W3196052105 creator A5030512993 @default.
- W3196052105 creator A5040427757 @default.
- W3196052105 creator A5079599342 @default.
- W3196052105 creator A5081322018 @default.
- W3196052105 date "2021-08-26" @default.
- W3196052105 modified "2023-09-27" @default.
- W3196052105 title "When should agents explore" @default.
- W3196052105 cites W1504212531 @default.
- W3196052105 cites W1522301498 @default.
- W3196052105 cites W1574672617 @default.
- W3196052105 cites W1591713425 @default.
- W3196052105 cites W1850488217 @default.
- W3196052105 cites W1863227302 @default.
- W3196052105 cites W2000514530 @default.
- W3196052105 cites W2015701831 @default.
- W3196052105 cites W2019377098 @default.
- W3196052105 cites W2034806191 @default.
- W3196052105 cites W2064675550 @default.
- W3196052105 cites W2098487995 @default.
- W3196052105 cites W2106289661 @default.
- W3196052105 cites W2109910161 @default.
- W3196052105 cites W2121863487 @default.
- W3196052105 cites W2134243390 @default.
- W3196052105 cites W2137864581 @default.
- W3196052105 cites W2145339207 @default.
- W3196052105 cites W2167489871 @default.
- W3196052105 cites W2168108292 @default.
- W3196052105 cites W2173564293 @default.
- W3196052105 cites W2310456331 @default.
- W3196052105 cites W2417786368 @default.
- W3196052105 cites W2556477470 @default.
- W3196052105 cites W2596982695 @default.
- W3196052105 cites W2616909309 @default.
- W3196052105 cites W2778689844 @default.
- W3196052105 cites W2786928559 @default.
- W3196052105 cites W2810754397 @default.
- W3196052105 cites W2899205164 @default.
- W3196052105 cites W2908064123 @default.
- W3196052105 cites W2914261249 @default.
- W3196052105 cites W2949273680 @default.
- W3196052105 cites W2949969799 @default.
- W3196052105 cites W2950416780 @default.
- W3196052105 cites W2950872548 @default.
- W3196052105 cites W2952908749 @default.
- W3196052105 cites W2962717849 @default.
- W3196052105 cites W2963276097 @default.
- W3196052105 cites W2963403143 @default.
- W3196052105 cites W2963477884 @default.
- W3196052105 cites W2963584407 @default.
- W3196052105 cites W2963615220 @default.
- W3196052105 cites W2964217298 @default.
- W3196052105 cites W2964751386 @default.
- W3196052105 cites W2971273232 @default.
- W3196052105 cites W2983170726 @default.
- W3196052105 cites W2987503309 @default.
- W3196052105 cites W2995102855 @default.
- W3196052105 cites W2995363872 @default.
- W3196052105 cites W2996695841 @default.
- W3196052105 cites W3013618273 @default.
- W3196052105 cites W3030493409 @default.
- W3196052105 cites W3042938692 @default.
- W3196052105 cites W3085832734 @default.
- W3196052105 cites W3091830134 @default.
- W3196052105 cites W3100329718 @default.
- W3196052105 cites W3126217842 @default.
- W3196052105 cites W3128571582 @default.
- W3196052105 cites W3132818773 @default.
- W3196052105 cites W3160722708 @default.
- W3196052105 cites W3168538470 @default.
- W3196052105 cites W3170969988 @default.
- W3196052105 cites W3123725896 @default.
- W3196052105 hasPublicationYear "2021" @default.
- W3196052105 type Work @default.
- W3196052105 sameAs 3196052105 @default.
- W3196052105 citedByCount "0" @default.
- W3196052105 crossrefType "posted-content" @default.
- W3196052105 hasAuthorship W3196052105A5016618355 @default.
- W3196052105 hasAuthorship W3196052105A5030512993 @default.
- W3196052105 hasAuthorship W3196052105A5040427757 @default.
- W3196052105 hasAuthorship W3196052105A5079599342 @default.
- W3196052105 hasAuthorship W3196052105A5081322018 @default.
- W3196052105 hasConcept C105795698 @default.
- W3196052105 hasConcept C107457646 @default.
- W3196052105 hasConcept C111472728 @default.
- W3196052105 hasConcept C127413603 @default.
- W3196052105 hasConcept C138885662 @default.
- W3196052105 hasConcept C154945302 @default.
- W3196052105 hasConcept C2776401178 @default.
- W3196052105 hasConcept C2776502983 @default.
- W3196052105 hasConcept C2780598303 @default.
- W3196052105 hasConcept C33923547 @default.
- W3196052105 hasConcept C41008148 @default.
- W3196052105 hasConcept C41895202 @default.
- W3196052105 hasConcept C48677424 @default.
- W3196052105 hasConcept C66938386 @default.
- W3196052105 hasConcept C67203356 @default.
- W3196052105 hasConcept C89611455 @default.