BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:A logical approach to data provenance. - James Cheney\, Informatic
 s\, University of Edinburgh
DTSTART:20061117T140000Z
DTEND:20061117T150000Z
UID:TALK5821@talks.cam.ac.uk
CONTACT:Tom Ridge
DESCRIPTION:While scientific computation historically has been synonymous 
 with large\nscale numerical computation and supercomputing\, scientists ar
 e now using\nincreasingly sophisticated computational techniques such as d
 atabases\nand decentralized "Grid" computation.  However\, scientific data
  is\nexpected to meet rigorous standards of data integrity\, and this is\n
 difficult to achieve using current tools.  One important ingredient of\nsc
 ientific data integrity is that data should be accompanied by\ndocumentati
 on of the process by which it was recorded: for example\,\ncreation/modifi
 cation timestamps\, descriptions of any operations\nperformed\, and identi
 ties of authors and intermediate sources.  This\ninformation is often call
 ed \\emph{provenance} or \\emph{lineage}.\n\nAlthough many provenance-trac
 king systems and provenance data models\nhave been proposed\, most of them
  are based on ad hoc definitions of\nprovenance\, some of which depend on 
 the syntax of the program (rather\nthan its semantics).  Thus\, the behavi
 or of such systems varies widely\,\nand we lack a uniform framework to com
 pare the correctness and\nexpressiveness of various approaches.  \n\nWe wi
 ll describe a new approach which is based on the idea that\nprovenance sho
 uld reflect the ways the output of a function depends on\nits input\; in p
 articular\, it should reflect counterfactual information\n(i.e.\, tell us 
 something about what would have happened if the input\nwere changed).  We 
 formalize this approach by defining a logic whose\nmodels are functions an
 d whose formulas are assertions about the\ndependence behavior of the func
 tion.  This provides a general approach\nto defining and reasoning about t
 he correctness of provenance-tracking\ntechniques.\n\nThis talk describes 
 joint work with Peter Buneman (U. Edinburgh)\, Stijn\nVansummeren (U. Hass
 elt\, Belgium)\, and Adriane Chapman (U. Michigan)\n\n
LOCATION:FW11
END:VEVENT
END:VCALENDAR
