PlayingWORDLE with UPPAAL STRATEGO

Peter G. Jensen; Kim G. Larsen; Marius Mikučionis

doi:10.1007/978-3-031-15629-8_15

PlayingWORDLE with UPPAAL STRATEGO

Peter G. Jensen, Kim G. Larsen, Marius Mikučionis^*

^*Kontaktforfatter

Publikation: Bidrag til bog/antologi/rapport/konference proceeding › Bidrag til bog/antologi › Forskning › peer review

Abstract

In this paper we model and solve the popular game Wordle using Uppaal Stratego. We model three different game-modes in terms of POMDPs, with more than 12,000 controllable actions. These constitute by far the largest models ever presented to Uppaal Stratego. Our experimental evaluation is encouraging: e.g. in the hard game-mode the partitioning-refinement learning method of Uppaal Stratego reduces the expected number of guesses from a baseline of 7.67 to 4.40 using 1 million training episodes. To better understand the convergence properties of our learning method we also study reduced versions of Wordle.

Originalsprog	Engelsk
Titel	A Journey from Process Algebra via Timed Automata to Model Learning : Essays Dedicated to Frits Vaandrager on the Occasion of His 60th Birthday
Redaktører	Nils Jansen, Mariëlle Stoelinga, Petra van den Bos
Antal sider	23
Forlag	Springer
Publikationsdato	2022
Sider	283-305
ISBN (Trykt)	978-3-031-15628-1
ISBN (Elektronisk)	978-3-031-15629-8
DOI	https://doi.org/10.1007/978-3-031-15629-8_15
Status	Udgivet - 2022

Navn	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Vol/bind	13560 LNCS
ISSN	0302-9743

Bibliografisk note

Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Adgang til dokumentet

10.1007/978-3-031-15629-8_15

AUB Link

Søg efter materialet i Aalborg Universitetsbiblioteks søgemaskine

Andre filer og links

Link to publication in Scopus

Citationsformater

Jensen, Peter G. ; Larsen, Kim G. ; Mikučionis, Marius. / PlayingWORDLE with UPPAAL STRATEGO. A Journey from Process Algebra via Timed Automata to Model Learning: Essays Dedicated to Frits Vaandrager on the Occasion of His 60th Birthday. red. / Nils Jansen ; Mariëlle Stoelinga ; Petra van den Bos. Springer, 2022. s. 283-305 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind 13560 LNCS).

@inbook{d75f2c7256db4cad9ebda83b6bed86c6,

title = "PlayingWORDLE with UPPAAL STRATEGO",

abstract = "In this paper we model and solve the popular game Wordle using Uppaal Stratego. We model three different game-modes in terms of POMDPs, with more than 12,000 controllable actions. These constitute by far the largest models ever presented to Uppaal Stratego. Our experimental evaluation is encouraging: e.g. in the hard game-mode the partitioning-refinement learning method of Uppaal Stratego reduces the expected number of guesses from a baseline of 7.67 to 4.40 using 1 million training episodes. To better understand the convergence properties of our learning method we also study reduced versions of Wordle.",

author = "Jensen, {Peter G.} and Larsen, {Kim G.} and Marius Miku{\v c}ionis",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.",

year = "2022",

doi = "10.1007/978-3-031-15629-8_15",

language = "English",

isbn = "978-3-031-15628-1",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "283--305",

editor = "Nils Jansen and Mari{\"e}lle Stoelinga and {van den Bos}, Petra",

booktitle = "A Journey from Process Algebra via Timed Automata to Model Learning",

address = "Germany",

}

Jensen, PG , Larsen, KG & Mikučionis, M 2022, PlayingWORDLE with UPPAAL STRATEGO. i N Jansen, M Stoelinga & P van den Bos (red), A Journey from Process Algebra via Timed Automata to Model Learning: Essays Dedicated to Frits Vaandrager on the Occasion of His 60th Birthday. Springer, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), bind 13560 LNCS, s. 283-305. https://doi.org/10.1007/978-3-031-15629-8_15

PlayingWORDLE with UPPAAL STRATEGO. / Jensen, Peter G.; Larsen, Kim G.; Mikučionis, Marius.
A Journey from Process Algebra via Timed Automata to Model Learning: Essays Dedicated to Frits Vaandrager on the Occasion of His 60th Birthday. red. / Nils Jansen; Mariëlle Stoelinga; Petra van den Bos. Springer, 2022. s. 283-305 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind 13560 LNCS).

Publikation: Bidrag til bog/antologi/rapport/konference proceeding › Bidrag til bog/antologi › Forskning › peer review

TY - CHAP

T1 - PlayingWORDLE with UPPAAL STRATEGO

AU - Jensen, Peter G.

AU - Larsen, Kim G.

AU - Mikučionis, Marius

PY - 2022

Y1 - 2022

N2 - In this paper we model and solve the popular game Wordle using Uppaal Stratego. We model three different game-modes in terms of POMDPs, with more than 12,000 controllable actions. These constitute by far the largest models ever presented to Uppaal Stratego. Our experimental evaluation is encouraging: e.g. in the hard game-mode the partitioning-refinement learning method of Uppaal Stratego reduces the expected number of guesses from a baseline of 7.67 to 4.40 using 1 million training episodes. To better understand the convergence properties of our learning method we also study reduced versions of Wordle.

AB - In this paper we model and solve the popular game Wordle using Uppaal Stratego. We model three different game-modes in terms of POMDPs, with more than 12,000 controllable actions. These constitute by far the largest models ever presented to Uppaal Stratego. Our experimental evaluation is encouraging: e.g. in the hard game-mode the partitioning-refinement learning method of Uppaal Stratego reduces the expected number of guesses from a baseline of 7.67 to 4.40 using 1 million training episodes. To better understand the convergence properties of our learning method we also study reduced versions of Wordle.

UR - http://www.scopus.com/inward/record.url?scp=85138154473&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-15629-8_15

DO - 10.1007/978-3-031-15629-8_15

M3 - Book chapter

AN - SCOPUS:85138154473

SN - 978-3-031-15628-1

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 283

EP - 305

BT - A Journey from Process Algebra via Timed Automata to Model Learning

A2 - Jansen, Nils

A2 - Stoelinga, Mariëlle

A2 - van den Bos, Petra

PB - Springer

ER -

Jensen PG , Larsen KG , Mikučionis M. PlayingWORDLE with UPPAAL STRATEGO. I Jansen N, Stoelinga M, van den Bos P, red., A Journey from Process Algebra via Timed Automata to Model Learning: Essays Dedicated to Frits Vaandrager on the Occasion of His 60th Birthday. Springer. 2022. s. 283-305. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind 13560 LNCS). doi: 10.1007/978-3-031-15629-8_15

PlayingWORDLE with UPPAAL STRATEGO

Abstract

Bibliografisk note

Adgang til dokumentet

AUB Link

Andre filer og links

Fingeraftryk

Citationsformater