expans search network borrow concept expand graph studi expans properti real world complex network social network unstructuredp peeror network extent properti exploit understand address problem decentr search produc sampl concis captur expans properti entir network collect refer expans signatur signatur find correspond magnitud maximum expans extent network effici search find evid standard graph theoret measur averag path length fail fulli explain level searchabl eas inform diffus dissemin network final demonstr high expans leverag facilit decentr search network show expans base search strategi outperform typic search method onlin sampl high central individu social network session introduct discov kinship small subset discov kinship small subset sampl commun structur sampl commun structur expans search network borrow concept expand graph studi expans properti real world complex network social network unstructuredp peeror network extent properti exploit understand address problem decentr search produc sampl concis captur expans properti entir network collect refer expans signatur signatur find correspond magnitud maximum expans extent network effici search find evid standard graph theoret measur averag path length fail fulli explain level searchabl eas inform diffus dissemin network final demonstr high expans leverag facilit decentr search network show expans base search strategi outperform typic search method optim model algorithm sibl reconstruct genet marker combinatori reconstruct half dsibl group microsatellit data full sibl group reconstruct microsatellit data studi problem reconstruct half sibl group studi theoret challeng computation demand problem paper present formul half sibl reconstruct problem prove apx hard present exact solut formul develop heurist biolog synthet dataset present experiment result compar lead altern softwar coloni show result competit half sibl group reconstruct presenc polygami preval natur approxim implicit cover problem biologi infer maximum likelihood hierarchi social network constant dfactor approxim algorithm identifi dynam commun propos approxim algorithm identifi commun dynam social network commun intuit character unusu dens knit subset social network notion problemat social interact chang time aggreg social network time radic misrepres exist chang commun structur recent propos optim base framework model dynam commun structur propos algorithm find structur base maximum weight bipartit match paper analyz perform guarante special case actor observ time instanc show algorithm small constant factor approxim optimum similar idea design approxim algorithm gener case individu possibl unobserv time show approxim factor increas twofold remain constant input size algorithm infer commun dynam network provabl approxim guarante demonstr gener algorithm real data set result confirm effici effect algorithm identifi dynam commun session introduct approxim cover pack problem paper approxim issu problem triangl pack full sibl reconstruct maximum profit coverag coverag gener special version set cover applic biologi rang fullsibl reconstruct wild popul biomolecular cluster paper show approxim properti differ consider inapproxim constant triangl pack problem improv previou result directli transform inapproxim gap astad problem maxim number satisfi equat set equat gf interest approxim result full sibl reconstruct problem answer question pose berger wolf al result maximum profit coverag problem match upper lower bound approxim ratio answer question pose hassin support nsf grant ii support nsf grant dbi ii ii mine period behavior dynam social network mine period behavior dynam social network impact structur predict diffus network find spread blocker dynam network abstract social interact conduit process spread popul rumor opinion behavior diseas context spread diseas undesir behavior import identifi blocker individu effect stop slow spread process popul problem resist systemat algorithm solut effort formul practic solut paper structur network measur indic blocker dynam social network contribut fold extend standard structur network measur dynam network compar block abil individu order rank dynam measur found simpl rank node static degre dynam version node degre perform consist surprisingli dynam cluster coeffici good indic static version perform wors random rank simpl practic local comput algorithm identifi kei blocker network structur predict tempor network frequent subgraph abstract type process model explicitli record interact set actor time applic common object seri observ predict interact occur futur present represent type tempor data gener adapt algorithm predict pattern interact arbitrari point futur test algorithm predict pattern email log correl stock close price social group herd plain zebra algorithm averag accuraci predict set interact unseen timestep knowledg algorithm predict interact finest time grain reconstruct sibl relationship wild popul accid birth make peopl sister brother make sibl mutual parentag maya angel technolog collect genotyp data natur popul open possibl investig fundament biolog phenomena includ behavior mate system herit adapt trait kin select dispers pattern power potenti genotyp inform rest abil reconstruct genealog relationship individu relationship includ parentag full half sibship higher order aspect pedigre area genealog infer parentag studi extens method pedigre infer kinship analysi exist make assumpt hold wild popul anim plant chapter focu full sibl relationship review exist method full sibship reconstruct microsatellit genet marker describ combinatori method sibl reconstruct base simpl mendelian law extens presenc error data describ gener consensu method combin sibl reconstruct result method present experiment comparison exist approach biolog simul data discuss rel merit drawback exist method suggest practic approach reconstruct sibl relationship wild popul framework commun identif dynam social network framework analysi dynam social network find pattern social interact popul wide rang applic includ diseas model cultur inform transmiss behavior ecolog social interact model network kei characterist social interact continu chang past analys social network essenti static inform time social interact place discard paper propos mathemat comput framework enabl analysi dynam social network explicitli make inform social interact occur onlin consensu agreement phylogenet tree abstract comput heurist primari method reconstruct phylogenet tree larg dataset larg scale phylogenet analys produc numer tree equival optim criteria heurist take signific amount time obtain optim tree simul experi biolog data score optim tree result heurist run fix long period time measur lack improv achiev unclear good criterion measur lack improv repres collect tree compact scientist monitor reconstruct progress consensu agreement tree common represent exist static algorithm produc tree increas lengthi comput time substanti paper present effici onlin algorithm comput strict major consensi maximum agreement subtre sharp bound bandwidth cliqu product index assign multichannel commun failur optim multichannel commun failur tanya berger wolf edward reingold octob abstract problem commun multipl channel channel fail inform channel lost goal recov origin inform redund analyz problem design encod minim decod error case channel failur give tight lower bound optim algorithm problem statement send lg bit inform independ channel partit lg bit subset subset receiv origin inform complet reconstruct trivial solut send block lg bit block fail arriv part inform lost altern send complet copi total lg bit channel 