User:Amanmehta/Application - Revision history

Amanmehta: /* Tasks completed till date */

2017-04-29T11:06:37Z

Tasks completed till date

Amanmehta: /* Workplan */

2017-04-07T10:04:22Z

Workplan

Amanmehta: /* Tasks completed till date */

2017-04-07T10:03:49Z

Tasks completed till date

Amanmehta: /* Detailed week-wise workplan */

2017-04-07T10:03:08Z

Detailed week-wise workplan

Amanmehta: /* Detailed week-wise workplan */

2017-04-07T09:52:33Z

Detailed week-wise workplan

Amanmehta: /* Detailed week-wise workplan */

2017-04-03T12:59:28Z

Detailed week-wise workplan

Amanmehta: /* Detailed week-wise workplan */

2017-04-03T12:37:05Z

Detailed week-wise workplan

Amanmehta: /* Tasks completed till date */

2017-04-03T10:35:50Z

Tasks completed till date

Amanmehta: /* Tasks completed till date */

2017-04-03T10:35:18Z

Tasks completed till date

Amanmehta: /* Tasks completed till date */

2017-04-03T10:13:28Z

Tasks completed till date

@@ Line 56: / Line 56: @@
 ***PER: ~84%
 **Current Status (ongoing):
-***WER: ~65% (~22% reduction)
+***WER: ~63% (~24% reduction)
-***PER: ~62% (~22% reduction)
+***PER: ~54% (~30% reduction)
 *Improved marathi bidix coverage for a wikipedia corpus (~3.6million tokens) from ~34% to ~49%. Reduced the count of unknown tokens by ~0.54m. (~1.9m still unknown).

@@ Line 70: / Line 70: @@
 *Target coverage~65%<br />
 For guj-hin pair:<br/>
-*Target WER<40% and better coverage.
+*Target WER <40%
+*apertium-guj dictionary coverage >70%

@@ Line 47: / Line 47: @@
 *Coding challenge:
 **Analysis:
-***Only 27% known tokens
+***Only 27% known tokens(improved to 74% now)
 ***About ~20% of unknown words are intransitive verbs
 ***About ~15% of unknown words are pronouns and their lexicals

@@ Line 93: / Line 93: @@
 *Analyse opportunity to improve dictionaries (tag editing/expand dictionaries)
 *Target: Bidix-coverage ~52%
 |-
 !colspan="2" style="text-align: right"|
@@ Line 116: / Line 114: @@
 *Bidix-coverage ~63%
 *Testvoc clean nouns
 *Start working with chunking (t2x)
 |-
 ! 2 !! June 6 - June 12
@@ Line 130: / Line 125: @@
 |-
 ! 3 !! June 13 - June 19
 |
 *Testvoc clean for adjectives, adverbs
 *Add lexical selection rules
@@ Line 138: / Line 131: @@
 **Bidix-coverage ~70%
 **WER <=25%
 |-
 ! 4 !!
 June 20 - June 26
 |
 *Finish with lexical selection rules and chunking.
 *Start working on CG
@@ Line 158: / Line 148: @@
 *Setup skeleton for mar-guj
 *Improve morphological analyzer if possible
 |-
 ! 6 !!
@@ Line 165: / Line 154: @@
 *Find good parallel corpora and add words in decreasing frequency in apertium-guj.
 **Coverage ~45%
 *Parallely improve coverage of mar-guj bilingual dictionary
 **Bidix-coverage ~30%
 *Guj-hin Bidix-coverage improvement
 |-
 ! 7 !! July 11 - July 17
 |
+*Work over a ~500 word story
-*Start adding verbs, adverbs, adjectives and pronouns.
-*Calculate new WER/PER.
+**Calculate WER, PER and document
+**Target WER <=55%
-*Reduce PER < 60%, document the results.
+*Even up nouns, pronouns
-*Start working on transfer grammar rules (t1x) using the common rules generated from post-edit analysis.
+*Even up for verbs, adjectives, adverbs
 |-
 ! 8 !! July 18 - July 24
 |
+*Testvoc clean for all classes
-*Evaluate the system and identify most frequent errors
+*Working on transfer grammar rules (t1x) using the common rules generated from post-edit analysis
-*Improve mar-guj bilingual dictionary adding to its coverage.
+**WER <=40%
-*Add more transfer rules for verbs, adjectives, adverbs and pronouns, Target:
+*Bidix-coverage ~50%
-**Coverage (~60%)
-**WER (<=30%)
 |-
 ! 9 !! July 25 - July  31
 |
+*Add transfer rules for nouns, pronouns
+*Add transfer rules for verbs, adjectives, adverbs.
-*Work over a ~500 word story
+*Start working on CG and disambiguation
-**Calculate WER, PER and document
-**Target WER <=55%
-*Even up nouns, pronouns
-*Even up for verbs, adjectives, adverbs
 |-
@@ Line 200: / Line 184: @@
 |
 *Continue working on disambiguation and its solutions.
+*Add required transfer/lexical selection rules to improve WER, PER.
-*Begin with chunking and t3x.
+*Begin with chunking and t3x
-*Add words and rules, starting with most frequent words (in decreasing frequency) for '''guj-hin pair'''.
 |-
 ! 11 !! August 7 - August 13
 |
+*Get another ~500 token story for guj-hin and improve WER.
-*Begin final testing for '''mar-guj pair'''
+**Target WER <=25%
-*Check for any major feedbacks to '''mar-hin pair''', work over it.
+*Regression testing for mar-guj pair
+*Evaluate test results, make the required changes, run tests again
 *User acceptance testing, gisting evaluation.
+*Mar-guj pair ready for or close to trunk and guj-hin improved. (Deliverable #2 and #3)
-*Start documentation and refactoring.
-*Final release: Mar-guj & mar-hin pairs ready for or close to trunk and guj-hin improved. ('''Deliverable #2 and #3''')
 |-
 ! 12 !! August 14 - August 21
 |
+*Regression testing for all the three pairs
-*Testing
+*Discuss with mentor about some final changes that must be made.
-*Detailed report on what further improvement could be made for the pairs, for future help to apertium.
+*Documentation, final release of all three pairs.
-*Evaluation and Refactoring.
+*Detailed analysis on what further improvement could be made for the pairs, for future help to apertium.
 |}

@@ Line 92: / Line 92: @@
 *Learn about the scope of lttoolbox and other relevant tools and get acquainted
 *Analyse opportunity to improve dictionaries (tag editing/expand dictionaries)
+*Target: Bidix-coverage ~52%
 |-
@@ Line 99: / Line 101: @@
 |
 *Read documents thoroughly
-*Study CG and rules for mar-hin.
 *Discuss strategies/solutions with mentor to tackle loopholes.
 *Add transfer rules for transitive and intransitive verbs.
@@ Line 105: / Line 106: @@
 *Test for (@,#,*) tokens
 *First regression testing.
-*Target: Coverage>50% (wikipedia corpus)
+*Target: Bidix-coverage ~60% (wikipedia corpus)
 |-
 ! 1 !! May 30 - June 5
 |
 *Even up nouns.
+*Add transfer rules for nouns, pronouns.
-*Testvoc clean for noun
 *Start working for pronouns, adverbs and adjectives
 *Add appropriate rules/stems.
+*Bidix-coverage ~63%
+*Testvoc clean nouns
 *Start working with chunking (t2x)
@@ Line 118: / Line 123: @@
 ! 2 !! June 6 - June 12
 |
+*Add transfer rules for adjectives, adverbs
 *Take another 500-word story.
 **Target: WER <50%
 *Post-edit translated texts. Analyze and look for common rules and add rules
+*Coverage ~67%
-*Testvoc clean for all classes
 |-
 ! 3 !! June 13 - June 19
@@ Line 129: / Line 133: @@
 |
+*Testvoc clean for adjectives, adverbs
-*Add more rules
+*Add lexical selection rules
-*Finish with t2x.
-*Get done with t3x.
 *Corpus test, measure improvement, targets:
-**Coverage ~67%
+**Bidix-coverage ~70%
-**WER <25%
+**WER <=25%
-*Read CG, lrx.
@@ Line 143: / Line 145: @@
 |
+*Finish with lexical selection rules and chunking.
-*Get most frequent mistakes at this stage
+*Start working on CG
-*If measurable improvement could be made, make required plan
+*Start working on disambiguation and its solutions
-*Run corpus testing to analyse to improvement. Target :
+*Refactoring and documentation.
-**Coverage ~70%
-**WER <= 20%
-*Run final tests, look for major mistakes and scope for corrections
 |-
 ! 5 !!
 June 27 - July 3
 |
+*Run corpus testing to analyse to improvement. Target :
-*Update documentation
+**Coverage ~70%
-*Run tests again, release this version ('''Deliverable #1''')
+**WER <= 20% ('''Deliverable #1''')
 *Setup skeleton for mar-guj
-*Start with expanding apertium-guj dictionary
 *Improve morphological analyzer if possible
@@ Line 163: / Line 163: @@
 July 4 - July 10
 |
-*Find good parallel corpora from wikipedia and add words and rules in decreasing frequency, especially in apertium-guj.
+*Find good parallel corpora and add words in decreasing frequency in apertium-guj.
+**Coverage ~45%
-**Continue for guj-hin bilingual dictionary expansion parallely.*
-*Even up nouns
+*Parallely improve coverage of mar-guj bilingual dictionary
-*Testvoc nouns
+**Bidix-coverage ~30%
+*Guj-hin Bidix-coverage improvement
 |-
@@ Line 187: / Line 189: @@
 |
+*Work over a ~500 word story
-*Testvoc cleanup for all classes
+**Calculate WER, PER and document
-*Work over a large corpora
+**Target WER <=55%
-*Regression and corpus testing
+*Even up nouns, pronouns
-*Work on CG and disambiguation
+*Even up for verbs, adjectives, adverbs
 |-

@@ Line 59: / Line 59: @@
 ***PER: ~62% (~22% reduction)
-*Improved marathi bidix coverage for a wikipedia corpus (~3.6million tokens) from 34% to 49%. Reduced the count of unknown tokens by ~0.54m. (~1.9m still unknown).
+*Improved marathi bidix coverage for a wikipedia corpus (~3.6million tokens) from ~34% to ~49%. Reduced the count of unknown tokens by ~0.54m. (~1.9m still unknown).
 == Workplan ==