Difference between revisions of "Daume et al, 2006"
From Cohen Courses
Jump to navigationJump to searchPastStudents (talk | contribs) |
PastStudents (talk | contribs) |
||
| Line 12: | Line 12: | ||
The key points of the paper are: | The key points of the paper are: | ||
| − | * a | + | * Authors state that SEARN is efficient, widely applicable, theoretically justified, and simple. |
| − | * | + | * SEARN looks at problems as a search problems, and learn classifiers that walk through the search space in a good way. |
| − | * | + | * Authors looked at 3 sample problems: [[AddressesProblem::Sequence Labeling]], [[AddressesProblem::Parsing]], and [[AddressesProblem::Machine Translation]] |
| − | * | + | * Efficacy of SEARN hinges on ability to compute an optimal/near-optimal policy. When optimal policy not available, authors suggest performing explicit search as an approximation. For segmentaiton and parsing, optimal policy is closed form; for summarization and machine translation, the optimal policy is not available. |
| + | |||
| + | == Example SEARN Usage == | ||
| + | |||
| + | '''Sequence Labeling''' | ||
| + | * Discussed POS-tagging and NP chunking | ||
| + | |||
| + | ''Tagging'' | ||
| + | * Produce label sequence from input sequence. Loss function is Hamming loss, search as left-to-right greedy search. | ||
| + | * Optimal Policy: | ||
| + | [[File:op-tagging.png]] | ||
| + | |||
| + | |||
| + | * ''Chunking'': Joint segmentation and labeling problem. Loss function is F1 measure | ||
| + | * Optimal Policy: | ||
| + | [[File:op-chunking.png]] | ||
| + | |||
| + | '''Parsing''' | ||
| + | * Looked at dependency parsing with shift-reduce framework | ||
| + | * ''Loss funtion'': Hamming loss over dependencies | ||
| + | * ''Decisions'': shift/reduce | ||
| + | * ''Optimal Policy'': | ||
| + | [[File:op-parsing.png]] | ||
| + | |||
| + | '''Machine Translation''' | ||
| + | * Frame as left-to-right translation problem. Search space over prefixes of translations, actions are adding a word (or phrase to end of existing translation) | ||
| + | * ''Loss function'': 1 - BLEU or 1 - NIST | ||
| + | * ''Optimal policy'': given set of reference translations R, English translation prefix e_1, ... e_i-1, what word (or phrase) should be produced next / are we finished. | ||
== Related papers == | == Related papers == | ||
* '''Search-based Structured Prediction''': This is the journal version of the paper that introduces the SEARN algorithm - [[RelatedPaper::Daume_et_al,_ML_2009]]. | * '''Search-based Structured Prediction''': This is the journal version of the paper that introduces the SEARN algorithm - [[RelatedPaper::Daume_et_al,_ML_2009]]. | ||
Revision as of 17:11, 30 September 2010
Citation
Duame, H., Langford, J., and Marcu, D. 2006. SEARN in Practice. Unpublished Manuscript
Online version
Summary
This unpublished manuscript describes how SEARN can be used for three Natural Language Processing related tasks: Sequence Labeling, Parsing, and Machine Translation
The key points of the paper are:
- Authors state that SEARN is efficient, widely applicable, theoretically justified, and simple.
- SEARN looks at problems as a search problems, and learn classifiers that walk through the search space in a good way.
- Authors looked at 3 sample problems: Sequence Labeling, Parsing, and Machine Translation
- Efficacy of SEARN hinges on ability to compute an optimal/near-optimal policy. When optimal policy not available, authors suggest performing explicit search as an approximation. For segmentaiton and parsing, optimal policy is closed form; for summarization and machine translation, the optimal policy is not available.
Example SEARN Usage
Sequence Labeling
- Discussed POS-tagging and NP chunking
Tagging
- Produce label sequence from input sequence. Loss function is Hamming loss, search as left-to-right greedy search.
- Optimal Policy:
- Chunking: Joint segmentation and labeling problem. Loss function is F1 measure
- Optimal Policy:
Parsing
- Looked at dependency parsing with shift-reduce framework
- Loss funtion: Hamming loss over dependencies
- Decisions: shift/reduce
- Optimal Policy:
Machine Translation
- Frame as left-to-right translation problem. Search space over prefixes of translations, actions are adding a word (or phrase to end of existing translation)
- Loss function: 1 - BLEU or 1 - NIST
- Optimal policy: given set of reference translations R, English translation prefix e_1, ... e_i-1, what word (or phrase) should be produced next / are we finished.
Related papers
- Search-based Structured Prediction: This is the journal version of the paper that introduces the SEARN algorithm - Daume_et_al,_ML_2009.
