Learning constraints through partial queries

Bessière, Christian; Carbonnel, Clément; Dries, Anton; Hébrard, Emmanuel; Katsirelos, George; Narodytska, Nina; Quimper, Claude-Guy; Stergiou, Kostas; Tsouros, Dimosthenis C.; Walsh, Toby

doi:10.1016/j.artint.2023.103896

Cited by 11 publications

(21 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

Section: Guiding Findscopementioning

confidence: 99%

“…The functions used in the literature (Bessiere et al 2013;Bessiere et al 2023) to find the scope of violated after a negative answer from the user (line 8 of Algorithm 1) work in a similar way. We will use FINDSCOPE from (Bessiere et al 2023) (shown in Algorithm 2) to demonstrate our method, but the same logic applies to all existing in the literature. FINDSCOPE methods recursively map the problem of finding a constraint to a simpler problem by removing blocks of variable assignments from the original query (the one asked in line 3 of Algorithm 1, to which the user answered "no") and asking partial queries to the user.…”

Section: Guiding Findscopementioning

confidence: 99%

“…However, no information about the violated constraints from B is used, and no guidance is utilized. The set of variables to remove, or keep, in the assignment is usually chosen randomly (Bessiere et al 2023).…”

Section: Guiding Findscopementioning

confidence: 99%

“…Now that we have formally formulated this problem, we can modify the constraints and/or add an objective function in order to improve the performance of FINDSCOPE. In order to achieve the logarithmic complexity from (Bessiere et al 2023), we can impose the following constraint in our CSP formulation:…”

Section: Guiding Findscopementioning

confidence: 99%

See 3 more Smart Citations

Learning to Learn in Interactive Constraint Acquisition

Tsouros,

Berden,

Guns

2024

AAAI

View full text Add to dashboard Cite

Constraint Programming (CP) has been successfully used to model and solve complex combinatorial problems. However, modeling is often not trivial and requires expertise, which is a bottleneck to wider adoption. In Constraint Acquisition (CA), the goal is to assist the user by automatically learning the model. In (inter)active CA, this is done by interactively posting queries to the user, e.g. does this partial solution satisfy your (unspecified) constraints or not. While interactive CA methods learn the constraints, the learning is related to symbolic concept learning, as the goal is to learn an exact representation. However, a large number of queries is required to learn the model, which is a major limitation. In this paper, we aim to alleviate this limitation by tightening the connection of CA and Machine Learning (ML), by, for the first time in interactive CA, exploiting statistical ML methods. We propose to use probabilistic classification models to guide interactive CA queries to the most promising parts. We discuss how to train classifiers to predict whether a candidate expression from the bias is a constraint of the problem or not, using both relation-based and scope-based features. We then show how the predictions can be used in all layers of interactive CA: the query generation, the scope finding, and the lowest-level constraint finding. We experimentally evaluate our proposed methods using different classifiers and show that our methods greatly outperform the state of the art, decreasing the number of queries needed to converge by up to 72%.

show abstract

Section: Guiding Findscopementioning

confidence: 99%

Section: Guiding Findscopementioning

confidence: 99%

Section: Guiding Findscopementioning

confidence: 99%

Section: Guiding Findscopementioning

confidence: 99%

See 2 more Smart Citations

Learning to Learn in Interactive Constraint Acquisition

Tsouros,

Berden,

Guns

2024

AAAI

View full text Add to dashboard Cite

show abstract

“…We also discuss some special tractable cases of the Juosan puzzle and pertinent mathematical analyses regarding the number of solutions. Investigation of tractable sub-problems and tractable variants of NP-complete problems are particularly important in theoretical computer science (see, e.g., [27,28]). Moreover, counting the number of solutions to a particular computational problem is also interesting from mathematical and computational perspectives, especially in counting complexity theory [29,30].…”

Section: Introductionmentioning

confidence: 99%

Note on Algorithmic Investigations of Juosan Puzzles

Ammar,

Arzaki,

Wulandari

2024

Jurnal Ilmu Komputer dan Informasi

View full text Add to dashboard Cite

We investigate several algorithmic and mathematical aspects of the Juosan puzzle—a one-player pencil-and- paper puzzle introduced in 2014 and proven NP-complete in 2018. We introduce an optimized backtracking technique for solving this puzzle by considering some invalid subgrid configurations and show that this algorithm can solve an arbitrary Juosan instance of size m × n in O(2mn) time. A C++ implementation of this algorithm successfully found the solution to all Juosan instances with no more than 300 cells in less than 15 seconds. We also discuss the special cases of Juosan puzzles of size m × n where either m or n is less than 3. We show that these types of puzzles are solvable in linear time in terms of the puzzle size and establish the upper bound for the number of solutions to the Juosan puzzle of size 1 × n. Finally, we prove the tractability of arbitrary m × n Juosan puzzles whose all territories do not have constraint numbers.

show abstract