ACL RD-TEC 1.0 Summarization of P06-1141
Paper Title:
AN EFFECTIVE TWO-STAGE MODEL FOR EXPLOITING NON-LOCAL DEPENDENCIES IN NAMED ENTITY RECOGNITION
AN EFFECTIVE TWO-STAGE MODEL FOR EXPLOITING NON-LOCAL DEPENDENCIES IN NAMED ENTITY RECOGNITION
Authors: Vijay Krishnan and Christopher D. Manning
Primarily assigned technology terms:
- algorithm
- approximate inference
- approximate randomization test
- belief propagation
- boundary detection
- computational linguistics
- computing
- conditional random field
- conditional random fields
- crfs
- cross validation
- cross-validation
- cutoff
- discriminative training
- encoding
- entity recognition
- entity recognition system
- error reduction
- exact inference
- gibbs sampling
- graphical model
- hmms
- learner
- learning
- loopy belief propagation
- maximum likelihood
- measuring
- modeling
- named entity recognition
- nlp
- perceptron
- randomization
- randomization test
- reasoning
- recognition
- recognition system
- reference resolution
- sampling
- sequence modeling
- tractable inference
- training procedure
- validation
- viterbi
- viterbi algorithm
- voted perceptron
Other assigned terms:
- 10-fold cross validation
- approach
- approximate randomization
- association for computational linguistics
- case
- contextual information
- convergence
- dependency graph
- development set
- distribution
- document
- document collection
- entity type
- entity types
- fact
- feature
- feature types
- hard constraint
- heuristic
- hypothesis
- labeling
- likelihood
- linguistics
- long-distance dependencies
- measure
- model structure
- named entities
- named entity
- names
- nlp tasks
- null hypothesis
- part-of-speech
- precision
- probabilistic information
- procedure
- relation
- relative error reduction
- running time
- sentence
- sentences
- sequence model
- statistical significance
- statistics
- technique
- test corpus
- test data
- test set
- text
- tokens
- training
- training data
- training set
- words
