This preview shows pages 1–13. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Weights on edges Definition 3: Weights on smoothed graph Structural Reranking Computing Graph Centrality Uniform Influx Weighted Influx Recursive Uniform Influx Recursive Weighted Influx Structural Reranking Incorporating initial scores Uniform Influx + LM Weighted Influx + LM , Recursive Uniform Influx + LM Recursive Weighted Influx + LM . Structural Reranking Generation Probabilities: Maximumlikelihood estimate (MLE) of w with respect to x Dirichletsmoothed version : Structural Reranking Generation Probabilities (cont.) KullbackLeibler divergence D Evaluation 8 8 0 8 8 8 4 6 3 7 4 64/96 7 5 10 7 29/48 UIn+LM > UIn = 10 WIn+Lm > WIn = 8 RUIn+LM > RUIn = 11 RWIn+LM > RWIn =10 39/48 Evaluation Links based on the vectorspace model Evaluation Nonstructural reranking Questions and Comments...
View Full
Document
 Spring '08
 DAVISON

Click to edit the document details