• Publications
  • Influence
Machine Learning of Temporal Relations
TLDR
This paper used temporal reasoning as an over-sampling method to dramatically expand the amount of training data, resulting in predictive accuracy on link labeling as high as 93% using a Maximum Entropy classifier on human annotated data.
Conditional Models of Identity Uncertainty with Application to Noun Coreference
TLDR
Several discriminative, conditional-probability models for coreference analysis are introduced, all examples of undirected graphical models that can incorporate a great variety of features of the input without having to be concerned about their dependencies.
SpatialML: Annotation Scheme, Corpora, and Tools
TLDR
In adapting the extent tagger to new domains, merging the training data from the above corpus with annotated data in the new domain provides the best performance.
Research Paper: Rapidly Retargetable Approaches to De-identification in Medical Records
TLDR
This paper describes a successful approach to de-identification that was developed to participate in a recent AMIA-sponsored challenge evaluation, and developed a method for tuning the balance of recall vs. precision in the Carafe system.
SpatialML: annotation scheme, resources, and evaluation
TLDR
In adapting the extent tagger to new domains, merging the training data from the ACE corpus with annotated data in the new domain provides the best performance.
An Integrated, Conditional Model of Information Extraction and Coreference with Appli
In a method and apparatus for making a pile-surfaced thermoplastic material by a tack-spin technique the laminate of backing web and adhered pile is hauled off from the heated drawing surface over a
Toward Conditional Models of Identity Uncertainty with Application to Proper Noun Coreference
TLDR
This paper introduces several discriminative, conditionalprobability models for coreference analysis, all examples of undirected graphical models that can incorporate a great variety of features of the input without having to be concerned about their dependencies.
MITRE system for clinical assertion status classification
TLDR
Using semantic attributes of concepts and information about document structure as features for statistical classification of assertions is a good way to leverage rule-based and statistical techniques in this task.
Three Approaches to Learning TLINKs in TimeML
TLDR
This paper focuses on the problem of learning temporal relations (called TLINKs) in TimeML, and draws attention to a bug in vector generation and the fact that the evaluation scheme was somewhat unrealistic.
...
1
2
3
4
5
...