[Ubuntu-BD] Bengali Machine Translation Engine

Sat Jan 31 16:16:33 GMT 2009

I am in fact joining the project of Anubadok by Golam Mortuza Hossain(
http://anubadok.sourceforge.net/index.htm). I am learning the technology.
And as a CSE student I should this out, what do you think? Not everyone came
in this world knowing everything.

And what I got in my little brain is:
1. Informing everyone who could contribute in this.
2. Reaching donors like ADB and UNDP to obtain funds for this.(Its huge so
needs lots of funding to settle it under one roof)
3. I feel all the past theses have to be taken into account. All the
research papers related to machine translator need to be gathered, listed
and indexed.
4. Need to call for more researches to obtain the clear path to accomplish
the task. Can be carried out by the different faculties in the Universities.
Students+Teachers of Bangla, Linguistics, Computer Science can help on these
research.
5. More thoughts can come from you...

On Sat, Jan 31, 2009 at 9:39 PM, Ashiqur Rahman Angel <angel at linux.org.bd>wrote:

> Great Lenin Vai. I don´t have any idea about a Machine Translation Engine,
> but looks like you have. So why don´t start planning about this project?
> Open a thread and tell us about your plan. Rest of us will glad to help you
> with this.
>
>
> On Sat, Jan 31, 2009 at 4:42 PM, 9el <lenin at phpxperts.net> wrote:
>
> > For these sort of projects UNDP provides aids I guess.
> >
> > @Shabab Mustafa : Instead of getting frustrated, we need to reach
> everyone.
> > Not everyone in these lists represent whole bangladesh/Bangalis. So, we
> > must
> > reach everyone. And in doing so we shall find the best guys to work with.
> >
> >
> > On Sat, Jan 31, 2009 at 2:44 PM, Shabab Mustafa <
> shabab.mustafa at gmail.com
> > >wrote:
> >
> > > Yes. That is one of the live examples that Jamil bhai gave. I was
> trying
> > to
> > > point out these kind of crisis situations that we already have observed
> > in
> > > some OS projects. I think, efficient project management is more
> important
> > > and comes first for potential contributor hunting and keep them alive
> to
> > do
> > > research, study, design work plan etc etc. That was and is my point of
> > > view.
> > >
> > >
> > > And I am having at a cautious stance because our past history says,
> what
> > > ever heats up fast, cools down faster.
> > >
> > >
> > >
> > > On Sat, Jan 31, 2009 at 1:50 PM, Jamil Ahmed <itsjamil at gmail.com>
> wrote:
> > >
> > > > Friends,
> > > >
> > > > Sorry for late response. I think we don't know what is out there
> > > > already. IIRC we have already mailed to different mailing lists about
> > > > ankur's english to bengali dictionary and machine translator project.
> > > > Very few people showed interest to work actively on it. Probably we
> > > > could not reach the potential contributors yet! :)
> > > >
> > > > Anyway, you can check the url below. Golam Mortuza Hossainm
> [gmhossain
> > > > at gmail dot com] is leading the project. If you are interested to
> > > > work/study you can mail him directly.
> > > >
> > > > http://anubadok.sourceforge.net/index.htm
> > > >
> > > > Cheers,
> > > > -Jamil
> > > >
> > > >
> > > > 2009/1/31 9el <lenin at phpxperts.net>:
> > > > > Thats what my point is. There's no reason to be afraid thinking
> that
> > > > > Automatic/Machine translation gonna be real hard job to accomplish.
> > And
> > > > > there's no saying Bangali people are worthless. And this is one of
> > the
> > > > main
> > > > > tasks Bangali people should do.
> > > > >
> > > > > Yes you may go like, "WHAT IS EVERYBODY'S TASK IS NOBODY'S TASK".
> We
> > > just
> > > > > need a plan. Bangali nation maybe a little slow learner. But
> everyone
> > > > > learns.
> > > > >
> > > > > Bangla being the most honored languages being the only language to
> > have
> > > > > martyrs. 21st February being regarded as the International
> > Mothertongue
> > > > Day
> > > > > are enough reasons to plead for this task to start this work as
> > > > immediately
> > > > > as possible. There can be no reason saying its too early to start!
> > What
> > > > we
> > > > > lack? point them together!
> > > > > What are our plus points? point them together!
> > > > >
> > > > > Be positive and contribute your positive thoughts related to this
> to
> > > > > everywhere and encourage, motivate all sorts of people for this.
> This
> > > > wont
> > > > > require 150 million Bangladeshi to accomplish the task. Take
> examples
> > > of
> > > > > other successful such works. Do comparative study. Its a doable
> task!
> > > > >
> > > > > And localization is nothing like machine translation at all.
> > > > > Localization/Internationalization is allotting the proper phrases
> for
> > > > preset
> > > > > phrases in English or other starter languages. And machine
> > translation
> > > is
> > > > > not a final task as well. If you use Google Translate you'll see
> that
> > > > they
> > > > > take your opinion on every translation made by the engine.
> > > > >
> > > > >
> > > > >
> > > > > On Sat, Jan 31, 2009 at 3:14 AM, tanjir <orko_147 at yahoo.com>
> wrote:
> > > > >
> > > > >> I somewhat agree with Shabab about the complexity of the project.
> > Ast
> > > > the
> > > > >> same time as Lenin vaiya is saying I also think we need to start.
> As
> > > it
> > > > is
> > > > >> not a test, and there is no harm in trying, we can definitely
> start.
> > > > There
> > > > >> might be failures, but hey! success never comes so easily!
> > > > >> After I saw the debate, I did a 15 minute research to find out how
> > > > really
> > > > >> this thing works. Most probably all of you are talking about
> Machine
> > > > >> Translation. Yahoo or Altavista's translation engine is called
> > Systran
> > > > which
> > > > >> won't be very effective in translating Bengali (even though they
> > claim
> > > > to
> > > > >> support Hindi). Another recent approach which is called
> Statistical
> > > > Machine
> > > > >> Translation (Google Translator) will solve the problems of
> > translating
> > > > from
> > > > >> rice to "chal" or "dhan" depending on the use and other issues
> > > (phrases,
> > > > >> proverbs, structural problems and so on).
> > > > >> OpenLogos or Apertium are the open source Machine Translation (not
> > > > >> statistical) projects available with source code (as usual). Those
> > who
> > > > are
> > > > >> interested can start from downloading the source codes and see how
> > > they
> > > > are
> > > > >> working for other languages.
> > > > >> I am not sure how many of you have access to IEEE and ACM digital
> > > > >> libraries. I found thousands of articles and journals on this
> topic.
> > > >  Also
> > > > >> there is a nice website with tons of information:
> > > > http://www.statmt.org/So I think instead of fighting on this topic,
> if
> > > we
> > > > can start researching
> > > > >> and organizing ourselves to start working, then it will give us
> more
> > > > >> jack-fruit (or mango- whatever you like) or we will end up being
> > > > >> another-jack-asses or jack-cows or jack monkeys. Don't know if it
> > > > helped;
> > > > >> just my 2 poisha.
> > > > >>
> > > > >> -----
> > > > >>
> > > > >> tanjir
> > > > >>
> > > > >> visit http://www.tanjir.net
> > > > >>
> > > > >> --- On Fri, 1/30/09, Shabab Mustafa <shabab.mustafa at gmail.com>
> > wrote:
> > > > >> From: Shabab Mustafa <shabab.mustafa at gmail.com>
> > > > >> Subject: Re: [Ubuntu-BD] [ot] digg like
> > > > >> To: "Ubuntu Bangladesh" <ubuntu-bd at lists.ubuntu.com>
> > > > >> Received: Friday, January 30, 2009, 3:40 PM
> > > > >>
> > > > >> >
> > > > >> > আপনার কথায় মনে হচ্ছে
> > > > >> সর্বস্তরের বাঙালি এতো
> > > > >> বৃহদায়তন প্রজেক্ট শুরু করার
> > > > >> মতো
> > > > >> > পরিপক্ক নয়। এটা উবুন্তু
> > > > >> বাংলাদেশ বা স্বেচ্ছাসেবকদের
> > > > >> প্রতি আহবান নয়। এটা
> > > > >> > কারো
> > > > >> > একার কাজ নয়। একার পক্ষে
> > > > >> সম্ভবও নয়। ১২ বছরের শিশুরা
> > > > >> অংশগ্রহণ করতে মানা নেই।
> > > > >> > সব পর্যায়ের সবাইকেই কীভাবে
> > > > >> এগুনো যায় ভাবতে হবে।
> > > > >> >
> > > > >> > আর যদি মনে করেন না এরা তো
> > > > >> পাঠানদের মতো পরাটা খায়না, তাই
> > > > >> পেস বোলিং সম্ভব নয়।
> > > > >> > তাহলে আলু পরাটা খেয়ে বাংলা
> > > > >> অনুবাদের উপায় শেখার
> > > > >> প্রশিক্ষণ নিতে পারে সবাই।
> > > > >> > কাদের কাছে নেবে বলুন তো?
> > > > >> আংরেজ? জার্মান? রুশ?
> > > > >>
> > > > >>
> > > > >> ভাইজান, আমার ধারণা আমরা একই
> > > > >> রাস্তায় হাঁটছি না। এখানে করতে
> > > > >> পারবে কি পারবে না
> > > > >> সে তর্ক কোথা থেকে আসছে? আবারও
> > > > >> বলছি আমার কথার মূল পয়েন্টা
> > > > >> হচ্ছে আগে আমাদের
> > > > >> কিছু ছোট-মাঝারি মানের
> > > > >> প্রজেক্ট চালিয়ে অভিজ্ঞতা
> > > > >> সঞ্চয় করে নিয়ে প্রস্তুত হয়ে
> > > > >> তারপর বড় কাজে হাত দেয়া দরকার।
> > > > >>
> > > > >> পেস বোলিং তো অনেকেই করে।
> > > > >> তাদের মধ্যে থেকে কয়জন
> > > > >> ম্যাকগ্রা, ওয়াসিম, ব্রেট লি
> > > > >> বের হয়? আর আপনার কি ধারণা এরা
> > > > >> বাংলা সিনেমার মত রাতে ঘুমিয়ে
> > > > >> ছিল আর সকালে উঠেই
> > > > >> দেখেছে দুনিয়া সেরা পেস বোলার
> > > > >> হয়ে গেছে?
> > > > >>
> > > > >> আবার পরোটা না খেলেও যেমন পেস
> > > > >> বোলার হওয়া যায় তেমনি পরোটা
> > > > >> খানেওয়ালা হলেই পেস
> > > > >> বোলার হয় না।
> > > > >> --
> > > > >> Ubuntu Bangladesh mailing list
> > > > >> ubuntu-bd at lists.ubuntu.com
> > > > >> https://lists.ubuntu.com/mailman/listinfo/ubuntu-bd
> > > > >>
> > > > >>
> > > > >>
> > > > >>
> > >  __________________________________________________________________
> > > > >> Looking for the perfect gift? Give the gift of Flickr!
> > > > >>
> > > > >> http://www.flickr.com/gift/
> > > > >> --
> > > > >> Ubuntu Bangladesh mailing list
> > > > >> ubuntu-bd at lists.ubuntu.com
> > > > >> https://lists.ubuntu.com/mailman/listinfo/ubuntu-bd
> > > > >>
> > > > > --
> > > > > Ubuntu Bangladesh mailing list
> > > > > ubuntu-bd at lists.ubuntu.com
> > > > > https://lists.ubuntu.com/mailman/listinfo/ubuntu-bd
> > > > >
> > > > --
> > > > Ubuntu Bangladesh mailing list
> > > > ubuntu-bd at lists.ubuntu.com
> > > > https://lists.ubuntu.com/mailman/listinfo/ubuntu-bd
> > > >
> > >
> > >
> > >
> > > --
> > > বাংলায় লিখুন, বাংলায় পড়ুন..
> > > WRITE IN BANGLA, READ IN BANGLA, VISIT http://omicronlab.com
> > > --
> > > Ubuntu Bangladesh mailing list
> > > ubuntu-bd at lists.ubuntu.com
> > > https://lists.ubuntu.com/mailman/listinfo/ubuntu-bd
> > >
> > --
> > Ubuntu Bangladesh mailing list
> > ubuntu-bd at lists.ubuntu.com
> > https://lists.ubuntu.com/mailman/listinfo/ubuntu-bd
> >
>
>
>
> --
> Angel
> GPG key: 0x34001F46
> Bangladesh Linux Users Alliance
> Fedora Ambassador Bangladesh
> http://fedoraproject.org/wiki/User:Angel
>
> Fedora -- Freedom² and rapid innovation
> --
> Ubuntu Bangladesh mailing list
> ubuntu-bd at lists.ubuntu.com
> https://lists.ubuntu.com/mailman/listinfo/ubuntu-bd
>