Dbrev dreaming of a database revolution
Download
1 / 13

DBrev: Dreaming of a Database Revolution - PowerPoint PPT Presentation


  • 105 Views
  • Uploaded on

DBrev: Dreaming of a Database Revolution. Gjergji Kasneci, Jurgen Van Gael, Thore Graepel Microsoft Research Cambridge, UK. Uncertainty in Applications. Intelligent data management with following requirements:. Store, represent, retrieve data. Assess accuracy and confidence.

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

PowerPoint Slideshow about ' DBrev: Dreaming of a Database Revolution' - claus


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
Dbrev dreaming of a database revolution

DBrev: Dreaming of a Database Revolution

Gjergji Kasneci, Jurgen Van Gael, Thore Graepel

Microsoft Research

Cambridge, UK


Uncertainty in applications
Uncertainty in Applications

Intelligent data management with following requirements:

  • Store, represent, retrieve data

  • Assess accuracyand confidence

  • Self diagnostic and calibration

+

DB & IR

Statistical ML


Main issues
Main Issues

Outrageous:

solve these problems simultaneously in integrated system…

 DBrev


Dbrev exploits large scale graphical model
DBrev Exploits Large-Scale Graphical Model

Combine logical constraints and sources of evidence about knowledge fragments into belief

network, e.g.:

Sample Belief Network for Aggregating User Feedback and Expertise on Knowledge Fragments,

Kasneci et al.: WSDM’11


Dbrev on information extraction and integration
DBrev on Information Extraction and Integration

Provenance through factor graphs in DBrev:


Dbrev on information extraction and integration1
DBrev on Information Extraction and Integration

Provenance through factor graphs in DBrev:

<MichaelJackson,

diedOn,

25-07-2009>

<MichaelJackson,

livesIn,

Ireland>

michaeljackson.com

f1’

f1

f2

michaeljackson-

sightings.com

wikipedia.org/wiki/Michael_Jackson



Dbrev on information extraction and integration3
DBrev on Information Extraction and Integration

Ambiguity & Context in DBrev:

Entity1

f

sameAs

f’

Ontological description/

Semantic features

Statistical fingerprint

derived from the Web

Entity

Entity2


Dbrev on information extraction and integration4
DBrev on Information Extraction and Integration

Consistency in DBrev:

<A, R, B> ^ <B, R, C> ^ <R, type, Transitive>  <A, R, C>

Extracted Triple: (“x”, “r”, “y”)

refersTo(“x”, A) ^ refersTo(“y”, C) ^ canBeDeduced(A, R, C)

 refersTo (“r”, R)


Dbrev on information extraction and integration5
DBrev on Information Extraction and Integration

Consistency in DBrev:

^

^

<A, R, B> ^ <B, R, C> ^ <R, type, Transitive>  <A, R, C>

Extracted Triple: (“x”, “r”, “y”)

v

refersTo(“x”, A) ^ refersTo(“y”, C) ^ canBeDeduced(A, R, C)

 refersTo (“r”, R)


Dbrev on information extraction and integration6
DBrev on Information Extraction and Integration

Retrieval & Discovery in DBrev:

partnerOf

locatedIn

Microsoft

$x

US

certifiedBy

SPARQL / Conjunctive Datalog / NAGA


Dbrev on information extraction and integration7
DBrev on Information Extraction and Integration

  • Approximate Matching

  • Entity / relationship similarity

  • Reasoning over relationship properties

  • Reasoning with temporal / spatial

  • constraints

Retrieval & Discovery in DBrev:

partnerOf

locatedIn

  • User Preference

  • Information needs

    • freshness, accuracy, popularity

  • Interests

    • context, background, current interest

Microsoft

$x

US

certifiedBy

SPARQL / Conjunctive Datalog / NAGA


Summary
Summary

DBrev builds on large-scale factor graph to simultaneously approach:

Retrieval &

Discovery

provenance

context

ambiguity

consistency

An inspiration to combine…

+

DB & IR

Statistical ML

… for the challenges ahead.


ad