"0"^^ . "21"^^ . "2008-01-01+01:00"^^ . . . . "2010-04-16+02:00"^^ . . "machine learning; constraint satisfaction"@en . . "LeCoS: merging machine LEarning and COnstraint Satisfaction"@en . "machine learning" . . "We propose to merge advanced techniques of two mature research fields: relational machine learning (RML) on one hand, and constraint satisfaction (CS) on the other hand. We have compelling reasons to believe that dramatic improvements in efficiency can be achieved by this fusion, on both the machine learning and the constraint satisfaction parts. This will likely result in the ability to solve real-world instances of problems that, due to computational demands, cannot be solved by traditional mutually isolated ML and CS algorithms. Specific subobjectives of the project are to exploit (i) CS techniques in subsumption tests and hypothesis consistency tests, routinnely performed in RML, and (ii) RML techniques for finding frequent structural patterns in CS instance descriptors, their solutions as well as in the computational processes leading to the solutions, and then use such patters as heuristic knowledge for solving subsequent CS instances."@en . . "Integrace strojov\u00E9ho u\u010Den\u00ED a spl\u0148ov\u00E1n\u00ED omezuj\u00EDc\u00EDch podm\u00EDnek" . "Navrhujeme integrovat pokro\u010Dil\u00E9 techniky dvou vysp\u011Bl\u00FDch v\u00FDzkumn\u00FDch obor\u016F, a to rela\u010Dn\u00EDho strojov\u00E9ho u\u010Den\u00ED (RSU) na jedn\u00E9 stran\u011B a spl\u0148ov\u00E1n\u00ED omezuj\u00EDc\u00EDch podm\u00EDnek (SOP) na stran\u011B druh\u00E9. M\u00E1me p\u0159esv\u011Bd\u010Div\u00E9 d\u016Fvody se domn\u00EDvat, \u017Ee t\u00EDmto spojen\u00EDm m\u016F\u017Eeme dos\u00E1hnout \u0159\u00E1dov\u00E9ho zv\u00FD\u0161en\u00ED efektivity jak na stran\u011B strojov\u00E9ho u\u010Den\u00ED, tak na stran\u011B spl\u0148ov\u00E1n\u00ED omezuj\u00EDc\u00EDch podm\u00EDnek.T\u00EDmto bude z\u0159ejm\u011B mo\u017Eno dos\u00E1hnout \u0159e\u0161en\u00ED skute\u010Dn\u00FDch instanc\u00ED\u00A0probl\u00E9m\u016F, kter\u00E9 jsou\u00A0vzhledem ke sv\u00FDm v\u00FDpo\u010Detn\u00EDm n\u00E1rok\u016Fm ne\u0159e\u0161iteln\u00E9 tradi\u010Dn\u00EDmi vz\u00E1jemn\u011B izolovan\u00FDmi algoritmy RSU a SOP. Konkr\u00E9tn\u00EDmi podc\u00EDli projektu je pou\u017E\u00EDt (i) technik SOP v procedur\u00E1ch testu subsumpce a testu konzistence hypot\u00E9zy\u00A0rutinn\u011B opakovan\u00FDch v RSU a (ii) technik RSU pro nalezen\u00ED \u010Dast\u00FDch strukturn\u00EDch vzor\u016F v deskriptorech instanc\u00ED probl\u00E9m\u016F SOP, jejich \u0159e\u0161en\u00ED a v\u00FDpo\u010Detn\u00EDch procesech\u00A0k t\u011Bmto \u0159e\u0161en\u00EDm vedouc\u00EDch, a tyto vzory vyu\u017E\u00EDt jako heuristickou znalost pro \u0159e\u0161en\u00ED dal\u0161\u00EDch instanc\u00ED SOP." . . . . "The goal of the project was to merge techniques of CSP and machine learning to improve (speed up, in particular) algorithms on both parts. This goal was achieved with significant impact in both fields. We consider the following achievements most important:- CSP for machine learning: both randomized and deterministic CSP techniques were leveraged in solving the problem of consistent clause search"@en . "http://www.isvav.cz/projectDetail.do?rowId=GA201/08/0509"^^ . . "2015-02-09+01:00"^^ . . . . . . "C\u00EDlem projektu bylo spojit techniky CSP a strojov\u00E9ho u\u010Den\u00ED, a t\u00EDmto spojen\u00EDm dos\u00E1hnout zlep\u0161en\u00ED (zejm. zrychlen\u00ED) algoritm\u016F na obou stran\u00E1ch. Tento c\u00EDl byl spln\u011Bn s v\u00FDznamn\u00FDm dopadem na oba obory. Za nejd\u016Fle\u017Eit\u011Bj\u0161\u00ED v\u00FDsledky pova\u017Eujeme:- CSP pro strojov\u00E9 u\u010Den\u00ED: uplatn\u011Bn\u00ED randomizovan\u00FDch i deterministick\u00FDch CSP technik pro \u00FAlohu vyhled\u00E1v\u00E1n\u00ED konzistentn\u00ED klauzule v induktivn\u00EDm logick\u00E9m programov\u00E1n\u00ED."@cs . "GA201/08/0509" . . . . "1"^^ . "2010-12-31+01:00"^^ . "1"^^ . "21"^^ .