Looking Forward: Project Plan

Version 0b

• ~20k papers
• Links to live (and cached) versions of papers
• Parsing both Postscript and PDF papers
• Deduplication of papers

Version 1/2

• ~40k papers
• Ref<->ref matching
• "Earliest" IE (and extracted metadata into XML representation of paper)
• Fielded search corresponding to the fields we're extracting
• Integration of various system pieces
• A concrete plan for Version 1
• Discussion of what kinds of queries we want

Version 1

• ~100k papers
• Web UI
• Ref segmentation
• Ref matching (both ref<->ref and ref<->paper)
• "Early IE" of headers and refs
• Fielded search including attributes of citing papers
• Display of both incoming and ongoing citations

Version 2

• ~500k papers
• Final IE
• Final UI
• Ref graph analysis
• Paper classification (by topic, etc.)

Version 3

• Authors as first-class objects

WebHome

-- JoshLewis - 03 Jul 2003
to top