Graph-based framework for multi-task multi-view learning

摘要

A system and method a Multi-Task Multi-View (M2TV) learning problem. The method uses the label information from related tasks to make up for the lack of labeled data in a single task. The method further uses the consistency among different views to improve the performance. It is tailored for the above complicated dual heterogeneous problems where multiple related tasks have both shared and task-specific views (features), since it makes full use of the available information.

主权项

1. A method for classifying entities from multiple channels in multi-task multi-view learning problems, wherein entities of different tasks are related with each other through shared or common features in multiple views, and a single learning task relating to a task specific feature in multiple views said method comprising:
generating a bi-partite graph-based model relating one or more entities and features in each said view; forming an objective function to impose consistency of each task and similarity constraints on common views of different tasks based on graphs generated from said model, wherein for each task, a first function g( ) is defined on entities which takes on class label values; and, a second function f( ) is defined on each view which takes values on the features in the view, said second function feature values used to determine the class label of an entity having such features; iteratively solving said objective function over each said task to obtain values for said first functions and second functions; and, generating labels that classify said entities based on obtained values for said first functions, wherein said entities include a plurality of candidate answers to questions posed in a question answering system operable in a first language, wherein a first task includes providing an answer to a question in said first language to a question posed, and a second task includes providing an answer to a question in a second language, different from said first language, in response to an identical question, wherein the task specific feature includes a language dependent feature, and the shared feature includes a language invariant feature of said answer in said first and second languages, wherein as programmed processor device is configured to perform one or more of said model generating, said forming, said iteratively solving and said label generating.