Training information extraction classifiers

A method, comprising:

• extracting, by a first information extraction technique, a first plurality of information objects from a natural language text;
• extracting, by a second information extraction technique, a second plurality of information objects from the natural language text;
• modifying values of one or more hyper-parameters of a conflict arbitration classifier;
• training the conflict arbitration classifier that produces a list of extracted information objects by performing a conflict arbitration function with respect to a set of conflicting information objects, such that a first information object of the set of conflicting information objects belongs to the first plurality of information objects a second information object of the set of conflicting information objects belongs to the second plurality of information objects;
• optimizing an information extraction quality metric reflecting quality of the list of extracted information objects, by iteratively repeating operations of extracting the first plurality of information objects, extracting the second plurality of information objects, modifying the values of the hyper-parameters, and training the conflict arbitration classifier.

A computer system, comprising:

• a memory;
• a processor, coupled to the memory, the processor configured to:
• train an information extraction classifier to extract a first plurality of information objects from a natural language text, wherein a definition of the information extraction classifier includes one or more hyper-parameters;
• produce a list of extracted information objects by performing a conflict arbitration function with respect to a set of conflicting information objects, such that a first information object of the set of conflicting information objects belongs to the first plurality of information objects and a second information object of the set of conflicting information objects belongs to a second plurality of information objects extracted from the natural language text;
• modify values of the hyper-parameters of the information extraction classifier;
• and
• optimize an information extraction quality metric reflecting quality of the list of extracted information objects, by iteratively repeating operations of training the information extraction classifier, performing the conflict arbitration function, and modifying the values of the hyper-parameters.

A computer-readable non-transitory storage medium comprising executable instructions that, when executed by a computer system, cause the computer system to:

• train an information extraction classifier to extract a first plurality of information objects from a natural language text, wherein a definition of the information extraction classifier includes one or more hyper-parameters;
• produce a list of extracted information objects by performing a conflict arbitration function with respect to a set of conflicting information objects, such that a first information object of the set of conflicting information objects belongs to the first plurality of information objects and a second information object of the set of conflicting information objects belongs to a second plurality of information objects extracted from the natural language text;
• modify values of the hyper-parameters of the information extraction classifier;
• and
• optimize an information extraction quality metric reflecting quality of the list of extracted information objects, by iteratively repeating operations of training the information extraction classifier, performing the conflict arbitration function, and modifying the values of the hyper-parameters.