User:AMR-KELEG/GSoC19 Proposal - Revision history

ScoopGracie: ScoopGracie moved page User:GSoC19 AMR-KELEG proposal to User:AMR-KELEG/GSoC19 Proposal: These go under individual user's pages

2020-04-27T15:30:37Z

ScoopGracie moved page User:GSoC19 AMR-KELEG proposal to User:AMR-KELEG/GSoC19 Proposal: These go under individual user's pages

AMR-KELEG: AMR-KELEG moved page User:AMR-KELEG to User:GSoC19 AMR-KELEG proposal: Move the proposal to a new document so that I can have my own personal page.

2020-04-23T20:00:46Z

AMR-KELEG moved page User:AMR-KELEG to User:GSoC19 AMR-KELEG proposal: Move the proposal to a new document so that I can have my own personal page.

AMR-KELEG: /* Work Plan */

2019-04-09T18:09:00Z

Work Plan

AMR-KELEG: /* Work Plan */

2019-04-09T18:06:28Z

Work Plan

AMR-KELEG: /* Why is it that you are interested in Apertium? */

2019-04-09T17:50:01Z

Why is it that you are interested in Apertium?

AMR-KELEG: /* Weighting of automata */

2019-04-09T04:20:39Z

Weighting of automata

AMR-KELEG: /* Contributions to Apertium */

2019-04-09T04:17:12Z

Contributions to Apertium

AMR-KELEG: /* How and who the project will benefit in society? */

2019-04-09T01:16:20Z

How and who the project will benefit in society?

AMR-KELEG: /* Coding challenge */

2019-04-04T20:50:19Z

Coding challenge

AMR-KELEG: /* Coding challenge */

2019-04-04T20:50:05Z

Coding challenge

@@ Line 118: / Line 118: @@
 |-
 | Week 1
-(27 May - 3 June)
+(27 May - 3 June) *
 | Develop the first supervised model (Unigram counts).
@@ Line 124: / Line 124: @@
 |-
 | Week 2
-(4 June - 10 June)
+(4 June - 10 June) *
 | Read, Understand and plan for implementing the publication for the first unsupervised model.
@@ Line 173: / Line 173: @@
 |
 |}
+* Note: I will have an examination as part of my Masters degree in June. I will inform my mentor whenever the schedule is announced. (I will need to study three days before the exam so my working hours will decrease during these days).
 == Contributions to Apertium ==

@@ Line 103: / Line 103: @@
 | Community Bonding
 | Communicate with the maintainers and get to know Apertium better.
 Solve some issues on Github.
 Prepare a better list of publications that are going to be implemented.
 Implement a baseline model for weigthing automata.
+Solve some issues on Github.
+Prepare a better list of publications that are going to be implemented.
+Implement a baseline model for weighting automata.
 |-
 | Week 1
 (27 May - 3 June)
 | Develop the first supervised model (Unigram counts).
 Write a shell script for generating weights using a tagged corpus.
 |-
@@ Line 115: / Line 126: @@
 (4 June - 10 June)
 | Read, Understand and plan for implementing the publication for the first unsupervised model.
+Document the findings in a blog post.
 |-
 | Week 3-4
-(11 June - 24 June)
+(11 June - 28 June)
 | Finalise the first unsupervised model and compare it to the supervised one.
 |-
@@ Line 127: / Line 140: @@
 (29 June - 5 July)
 | Read, Understand and plan for implementing the publication for the second unsupervised model.
+Document the findings in a blog post.
 |-
 | Week 6
@@ Line 135: / Line 150: @@
 (13 July - 22 July)
 | Read, Understand and plan for implementing the publication for the second unsupervised model.
+Document the findings in a blog post.
 |-
 | Week 8
-(23 July - 12 July)
+(23 July - 26 July)
-| Implement the second unsupervised model.
+| Implement the third unsupervised model.
 |-
 | '''Evaluation 2'''
 '''Deliverables: A shell script for using the second unsupervised model and a plan for implementing the third one.'''
 |-
-| Week 9
+| Week 9-10
-(27 July - 2 August)
+(27 July - 9 August)
-| Implement the third unsupervised model.
-|-
-| Week 10
-(3 August - 9 August)
 | Solve issues related to the developed models.
+Finalise and package the scripts.
 |-
 | Week 11-12

@@ Line 49: / Line 49: @@
 = Project Information =
 == Why is it that you are interested in Apertium? ==
-I am interested in NLP and especially the idea of how to enable machines to understand and reason about human languages. This field have made it possible to perform tasks that couldn't have been done before. One of the interesting applications of NLP is Machine Translation. Machine Translation programs such as: Apertium have permitted people to automatically translate text from other languages. This has improved the way that we people share knowledge and experience.
+I am interested in NLP and especially the idea of how to enable machines to understand and reason about human languages. This field have made it possible to perform tasks that couldn't be done before. One of the interesting applications of NLP is Machine Translation. Machine Translation programs such as: Apertium have permitted people to automatically translate text from other languages. This has improved the way that we people share knowledge and experience.
 One of the main points that attracted me to Apertium is the fact that most of the maintainers are actually researchers. So not only the program is developed by experienced and skilled developers but also it's maintained by academic researchers that have good understanding of the field and the limitations/difficulties of automatic machine translation.

@@ Line 83: / Line 83: @@
 == Relevant publications ==
 === Weighting of automata ===
-* Parameter Estimation for Probabilistic Finite-State Transducers (https://aclweb.org/anthology/P02-1001):
+* Parameter Estimation for Probabilistic Finite-State Transducers (https://aclweb.org/anthology/P02-1001)
-** Main points:
 === Background papers ===

@@ Line 161: / Line 161: @@
 == Contributions to Apertium ==
-I have managed to fixed multiple issues in different repositories.
+I have managed to fix multiple issues in different repositories of Apertium's code-base.
 Merged pull requests:

← Older revision	Revision as of 15:30, 27 April 2020
(No difference)

← Older revision	Revision as of 20:00, 23 April 2020
(No difference)

@@ Line 63: / Line 63: @@
 The project should help developers and users generate weights for FST in an unsupervised way. This has the advantage of using large corpus to generate these weights without the requirement of manual annotation.
+Developers and users aiming at using Apertium to translate documents whose vocabulary is too specific to be included in the available annotated datasets will be able to adjust the weights of the FST according to a large un-annotated corpus of relevant documents.
-{{comment|And how does this in turn benefit society? —[[User:Firespeaker|Firespeaker]] ([[User talk:Firespeaker|talk]]) 19:57, 30 March 2019 (CET)}}
 == Coding challenge ==

@@ Line 76: / Line 76: @@
 * Use hfst-strings2fst and hfst-fst2txt to convert the weighted FST into att format.
 * Use lt-comp to generate a bin file for the FST using apertium's lttoolbox.
-* Use lt-proc using the new weighted fst.
+* Use lt-proc given the new weighted fst.
 Note: You will need to build the master branch of lttoolbox so that the analyses weights are computed correctly.