Scipediacontent: Scipediacontent moved page Draft Content 866252350 to Finkel et al 2010a

2021-02-02T03:12:51Z

Scipediacontent moved page Draft Content 866252350 to Finkel et al 2010a

Scipediacontent: Created page with " == Abstract == The end-to-end performance of natural language processing systems for compound tasks, such as question answering and textual entailment, is often hampered by..."

2021-02-02T03:12:44Z

Created page with " == Abstract == The end-to-end performance of natural language processing systems for compound tasks, such as question answering and textual entailment, is often hampered by..."

New page

== Abstract ==

The end-to-end performance of natural language processing systems for compound tasks, such as question answering and textual entailment, is often hampered by use of a greedy 1-best pipeline architecture, which causes errors to propagate and compound at each stage. We present a novel architecture, which models these pipelines as Bayesian networks, with each low level task corresponding to a variable in the network, and then we perform approximate inference to find the best labeling. Our approach is extremely simple to apply but gains the benefits of sampling the entire distribution over labels at each stage in the pipeline. We apply our method to two tasks -- semantic role labeling and recognizing textual entailment -- and achieve useful performance gains from the superior pipeline architecture.

== Original document ==

The different versions of the original document can be found in:

* [http://dl.acm.org/ft_gateway.cfm?id=1610162&type=pdf http://dl.acm.org/ft_gateway.cfm?id=1610162&type=pdf]

* [http://dx.doi.org/10.3115/1610075.1610162 http://dx.doi.org/10.3115/1610075.1610162]

* [https://nlp.stanford.edu/jrfinkel/papers/pipeline.pdf https://nlp.stanford.edu/jrfinkel/papers/pipeline.pdf],
: [http://www.robotics.stanford.edu/~ang/papers/emnlp06-bayesianpipeline.pdf http://www.robotics.stanford.edu/~ang/papers/emnlp06-bayesianpipeline.pdf],
: [http://ai.stanford.edu/~ang/papers/emnlp06-bayesianpipeline.pdf http://ai.stanford.edu/~ang/papers/emnlp06-bayesianpipeline.pdf],
: [http://www-nlp.stanford.edu/~manning/papers/pipeline.pdf http://www-nlp.stanford.edu/~manning/papers/pipeline.pdf],
: [https://www.aclweb.org/anthology/W06-1673.pdf https://www.aclweb.org/anthology/W06-1673.pdf],
: [https://www.aclweb.org/anthology/W06-1673 https://www.aclweb.org/anthology/W06-1673],
: [https://dblp.uni-trier.de/db/conf/emnlp/emnlp2006.html#FinkelMN06 https://dblp.uni-trier.de/db/conf/emnlp/emnlp2006.html#FinkelMN06],
: [https://dl.acm.org/citation.cfm?id=1610075.1610162 https://dl.acm.org/citation.cfm?id=1610075.1610162],
: [https://academic.microsoft.com/#/detail/2083195487 https://academic.microsoft.com/#/detail/2083195487]

← Older revision	Revision as of 03:12, 2 February 2021
(No difference)

Finkel et al 2010a - Revision history

Scipediacontent: Scipediacontent moved page Draft Content 866252350 to Finkel et al 2010a

Scipediacontent: Created page with " == Abstract == The end-to-end performance of natural language processing systems for compound tasks, such as question answering and textual entailment, is often hampered by..."