learn commun relat reinforc learn relat reinforc learn rrl emerg machin learn commun promis subfield reinforc learn rl upgrad rl techniqu relat represent state action learn valuefunct frequent subgraph mine outerplanar graph recent year increas interest frequent pattern discoveri larg databas graph structur object frequent connect subgraph mine problem tree dataset solv increment polynomi time intract arbitrari graph databas exist approach resort heurist strategi restrict search space identifi practic relev tractabl graph class tree paper class outerplanar graph strict gener tree develop frequent subgraph mine algorithm outerplanar graph show work increment polynomi time practic relev subclass behav outerplanar graph polynomi simpl cycl evalu algorithm empir chemo bioinformat applic predict clinic condit coronari bypass surgeri dynam data analysi comparison prune criteria probabl tree effici frequent connect subgraph mine graph bound tree dwidth effici frequent connect subgraph mine graph bound tree dwidth proceed acm sigkdd workshop statist relat learn bioinformat learn commun relat reinforc learn relat reinforc learn rrl emerg machin learn commun promis subfield reinforc learn rl upgrad rl techniqu relat represent state action learn valuefunct mont dcarlo tree search poker expect reward distribut poker plai comput bot divid categori game theoret bot plai strategi rise nash equilibrium bot imposs beat exploit optim oppon type bot exploit bot emploi game tree search oppon model techniqu discov exploit weak handl miss valu censor data pca pharmacolog matric dynam data analysi data mine predict clinic stabil polynomi ddelai enumer monoton graph class deriv distanc metric gener relat strebio statist relat learn mine bioinformat bioinformat applic domain inform natur repres term relat heterogen object modern experiment data acquisit techniqu studi complex interact biolog system rais interest challeng machin learn data mine research amount data huge inform observ measur noisi report present review acm sigkdd workshop statist relat learn mine bioinformat strebio held pari june aim workshop provid forum share challeng result idea frontier field statist relat learn field bioinformat bay drelat learn oppon model incomplet inform dlimit poker propos oppon model approach nolimit texa hold em poker start learn prior gener expect oppon behavior learn relat regress tree function adapt prior specif oppon import asset approach learn incomplet inform know player hand train game activ learn high throughput screen effici comput graph dbase metric classif small molecul decis tree answer network tempor differ dnetwork state represent intellig agent continu challeng abstract unavoid larg state space predict represent offer obtain state abstract replac state set predict futur interact effici frequent connect subgraph mine graph bound treewidth anti dmonoton overlap dgraph support measur learn direct probabilist logic model order dsearch versu structur dsearch abstract discuss learn recurs direct probabilist logic model relat data problem tackl upgrad structur search algorithm initi propos bayesian network paper propos upgrad algorithm order search bayesian network found work structur search experiment compar upgrad algorithm relat domain conclud signific differ algorithm term qualiti learnt model order search significantli faster gener order dsearch learn direct probabilist logic model abstract recent increas interest direct probabilist logic model varieti languag describ model propos author provid high level argument show principl model languag learn data propos learn algorithm studi detail introduc algorithm gener order search learn structur condit probabl distribut cpd direct probabilist logic model algorithm upgrad order search algorithm bayesian network relat probabl tree represent cpd present experi block world domain gene domain cora dataset polici learn restrict polici space learn direct probabilist logic model order dsearch versu structur dsearch abstract discuss learn recurs direct probabilist logic model relat data problem tackl upgrad structur search algorithm initi propos bayesian network paper propos upgrad algorithm order search bayesian network found work structur search experiment compar upgrad algorithm relat domain conclud signific differ algorithm term qualiti learnt model order search significantli faster transfer learn reinforc learn problem partial polici recycl abstract paper investig relat transfer learn reinforc learn function approxim supervis learn concept drift present increment relat regress tree algorithm capabl deal concept drift tree restructur show enabl reinforc learner precis learner transfer knowledg task recycl part gener function hold interest inform task illustr perform algorithm experi supervis learn task concept drift reinforc learn task transfer knowledg easier relat task onlin learn exploit relat model reinforc learn recent year grow interest rich represent relat languag reinforc learn express languag advantag term gener reason extend exist approach relat set trivial problem paper present step onlin learn exploit relat model propos represent transit reward function learn onlin present method exploit model augment relat reinforc learn algorithm plan techniqu benefit robust approach evalu experiment learn direct probabilist logic model order dsearch polynomi dtime metric outerplanar graph gener order dsearch learn direct probabilist logic model abstract recent increas interest direct probabilist logic model varieti languag describ model propos author provid high level argument show principl model languag learn data propos learn algorithm studi detail introduc algorithm gener order search learn structur condit probabl distribut cpd direct probabilist logic model algorithm upgrad order search algorithm bayesian network relat probabl tree represent cpd present experi block world domain gene domain cora dataset remauv relat model tree learner frequent subgraph mine outerplanar graph recent year increas interest frequent pattern discoveri larg databas graph structur object frequent connect subgraph mine problem tree dataset solv increment polynomi time intract arbitrari graph databas exist approach resort heurist strategi restrict search space identifi practic relev tractabl graph class tree paper class outerplanar graph strict gener tree develop frequent subgraph mine algorithm outerplanar graph show work increment polynomi time practic relev subclass behav outerplanar graph polynomi simpl cycl evalu algorithm empir chemo bioinformat applic frequent subgraph mine outerplanar graph recent year increas interest frequent pattern discoveri larg databas graph structur object frequent connect subgraph mine problem tree dataset solv increment polynomi time intract arbitrari graph databas exist approach resort heurist strategi restrict search space identifi practic relev tractabl graph class tree paper class outerplanar graph strict gener tree develop frequent subgraph mine algorithm outerplanar graph show work increment polynomi time practic relev subclass behav outerplanar graph polynomi simpl cycl evalu algorithm empir chemo bioinformat applic refin aggreg condit relat learn relat learn predict individu base properti properti set relat individu system aggreg summar set featur introduc compar result aggreg function threshold case set aggreg gener complex queri present framework refin complex aggreg condit dimens aggreg function queri gener set threshold propos aggreg refin oper effici search hypothesi space benefici relat learner aggreg applic implement refin oper relat decis tree induct system experiment result show signific effici gain comparison advanc refin oper graph kernel gaussian process relat reinforc learn relat reinforc learn learn techniqu relat state action space aim enabl agent learn act environ natur represent tupl constant case learn algorithm approxim map state action pair call ualiti reliabl handl relat represent state action pair paper investig logic bayesian network relat probabilist logic model review logic bayesian network languag probabilist logic model discuss relat probabilist relat model bayesian logic program probabilist logic model probabilist logic model model combin aspect probabl theori aspect logic program order logic relat languag recent varieti languag describ model introduc languag techniqu exist learn model data exampl probabilist relat model prm bayesian logic program blp languag popular relat data mine commun introduc languag logic bayesian network lbn strongli relat prm blp solv problem respect knowledg represent relat express intuit prm blp lbn follow principl knowledg base model construct offer languag gener probabilist logic knowledg provid methodolog construct proposit model base knowledg specif comparison approach learn probabl tree probabl tree probabl estim tree pet decis tree probabl distribut leav altern approach learn probabl tree propos comparison approach exist logic bayesian network relat probabilist logic model review logic bayesian network languag probabilist logic model discuss relat probabilist relat model bayesian logic program probabilist logic model probabilist logic model model combin aspect probabl theori aspect logic program order logic relat languag recent varieti languag describ model introduc languag techniqu exist learn model data exampl probabilist relat model prm bayesian logic program blp languag popular relat data mine commun introduc languag logic bayesian network lbn strongli relat prm blp solv problem respect knowledg represent relat express intuit prm blp lbn follow principl knowledg base model construct offer languag gener probabilist logic knowledg provid methodolog construct proposit model base knowledg specif multi dagent relat reinforc learn abstract paper studi relat reinforc learn multi agent set grow evid reinforc learn research commun relat represent state space benefit proposit complex task plan inform retriev web repres natur relat form relat structur exploit multi agent reinforc learn task studi singl agent context paper attempt bridg gap relat reinforc learn rrl multi agent system ma precis explor relat structur state space multi agent reinforc learn context search compound goal relev zone game condens represent induct logic program mine frequent datalog queri queri cover exampl equival redund equival due data set regular background theori avoid gener redund claus introduc type condens represent specif introduc free close claus defin data set semant free close claus account logic background theori algorithm emploi represent present experiment evalu number benchmark problem induct logic program compact represent knowledg base induct logic program applic induct logic program ilp learn occur knowledg base larg number exampl store knowledg base relat instanc base regress relat reinforc learn full paper topic appear proceed twentieth intern graph kernel gaussian process relat reinforc learn relat reinforc learn learn techniqu relat state action space aim enabl agent learn act environ natur represent tupl constant case learn algorithm approxim map state action pair call ualiti reliabl handl relat represent state action pair paper investig compact represent knowledg base ilp induct system includ ilp system learn knowledg base structur exampl practic situat center represent lot redund cluster instanc base learn order logic improv effici induct logic program queri pack induct logic program relat learn power paradigm machin learn data mine order ilp practic effici ilp system improv substanti end notion queri pack introduc structur set similar queri mechan execut queri pack complex analysi show consider effici improv achiev queri pack execut mechan claim support empir result obtain incorpor support queri pack execut exist learn system speed relat reinforc learn increment order decis tree learner relat reinforc learn rrl learn techniqu combin standard reinforc learn induct logic program enabl learn system exploit structur knowledg applic domain polynomi time comput metric point set measur similar distanc set point metric space import problem machin learn applic disciplin comput geometri philosophi scienc method updat chang theori recent eiter mannila propos measur comput polynomi time distanc function mathemat sens satisfi triangl inequ learn heurist tild execut queri pack ilp induct logic program system send larg number queri databas lattic structur queri typic select queri highli similar consequ independ execut queri involv lot redund comput propos mechan execut hierarch structur set queri queri pack lot redund comput remov incorpor queri pack execut mechan ilp system tild warmr implement prolog engin ilprolog support pack execut lower level experiment result demonstr signific effici gain queri pack execut mechan gener natur incorpor ilp system similar effici improv expect belief network neutral depend conceptu cluster belief network neutral depend conceptu cluster top ddown induct cluster tree approach cluster present adapt basic top induct decis tree method cluster aim emploi principl instanc base learn result methodolog implement tic top induct cluster tree system order cluster tic system emploi order logic decis tree represent induct logic program system tild experi tic present proposit relat domain gener refin oper learn prenex conjunct normal form induct logic program consid exclus univers quantifi theori add express gener prenex conjunct normal form existenti variabl learn refin oper learn method ilp extend refin oper pcnf extend substitut pcnf usual substitut gener formula special existenti variabl articl defin substitut special pcnf weakli complet downward refin oper base oper implement simpl learn system pcl type pcnf instanc base function learn principl instanc base function learn present ibfl set posit exampl function predic exampl true ground fact illustr input output behaviour predic purpos predict output predic input assumpt background theori input output predic consist structur term ibfl techniqu address problem combin idea instanc base learn order distanc analog case base reason argu ibfl handl complex deepli nest term present techniqu isol compon larger system deal logic languag learn challeng top ddown induct cluster tree approach cluster present adapt basic top induct decis tree method cluster aim emploi principl instanc base learn result methodolog implement tic top induct cluster tree system order cluster tic system emploi order logic decis tree represent induct logic program system tild experi tic present proposit relat domain framework defin distanc dorder logic object paper develop framework distanc claus distanc model framework parametris measur distanc atom take account subterm common distinct atom set atom measur distanc set constant number variabl complex distanc comput polynomi bound size object initi experi show framework basi good cluster algorithm framework consist level level choos distanc atom level upgrad distanc distanc set atom propos framework generalis polynomi time comput similar measur propos eiter mannila instanc real distanc function comput polynomi time develop binari prototyp function set point prototyp fun 