date : 2014/09/18 author : niket tandon, gerard de melo, fabian suchanek, gerhard weikum source :...
TRANSCRIPT
Query-Performance prediction: Setting the Expectations Straight
Date : 2014/09/18Author : Niket Tandon , Gerard de Melo , Fabian Suchanek , Gerhard WeikumSource : WSDM14Advisor : Jia-ling KohSpeaker : Shao-Chun Peng1WebChild: Harvesting and Organizing Commonsense Knowledge from the WebOutline2IntroductionApproachExperimentalConclusion
Introduction3What is knowledge bases ?Web searchtext analyticsRecommendations in social media
Introduction4MotivationComputers completely lack this kind of commonsense knowledge
Round and red
Introduction5How about teach computer?
Round and red
Introduction6PurposeAutomatically extracting and cleaning commonsense properties from the Web
Introduction7salient characteristics and unique qualitiesFine-grained assertionsDisambiguated argumentsMinimal supervisionOutline8IntroductionApproachRange PopulationDomain PopulationComputing AssertionsExperimentalConclusion
Sub task9Range PopulationDomain PopulationComputing Assertions
Range Population10Candidate GatheringGraph ConstructionEdgeWeightingLabel Propagation (LP)hasTaste: {delicious, spicy, hot, sweet, etc...}Range Population11Candidate GatheringGoogle N-gram corpus(5-gram)checking for the presence of the word r , any of its synonyms40000 adj. for all relation
Range Population12Graph Construction3 kinds of edgeedges among wordsedges between two senses u-ai and w-aj edges between words and senses
Range Population13EdgeWeightingedges among words
edges between two senses u-ai and w-aj taxonomic relatedness within WordNet [17]edges between words and sensesthe sense frequencies as a basis for edge weights
glosses of their respective hyponyms and hypernyms13Range Population14Label Propagation (LP)
ABDCABDCRelaton:0.8Dummy:0.2Relaton:0.7Dummy:0.30.80.2Relaton:0.8Dummy:0.20.80.2Relaton:0.1Dummy:0.9Domain Population15Candidate Gathering*Graph ConstructionEdgeWeightingLabel Propagation (LP)Beef: hasTaste
sour-a2:the taste experience when vinegar or lemon juice is taken . . (vinegar, sour-a2) and (lemon juice, sour-a2)gloss for sour-a2 reads the taste experience when vinegar or lemon juice is taken . . . . We generatetwo assertions from this: (vinegar, sour-a2) and (lemon juice, sour-a2)15Computing Assertions
16Graph Construction*EdgeWeighting*Label Propagation (LP)
Outline17IntroductionApproachExperimentalConclusion
Range Population18
Domain Population19
Computing Assertions20
Outline21IntroductionApproachExperimentalConclusion