[Haskell-cafe] Haskell-Cafe Digest, Vol 202, Issue 14

Michal J Gajda mgajda at mimuw.edu.pl
Sat Jun 13 12:27:29 UTC 2020


After careful examination of the paper,
I inferred that it mostly implemented simple loop translation,
and worked only as long as the problem was small and easy to refine
by direct translation token by token.
It also had problems with preserving semantics of these simple loops.
So while I am certain that Facebook AI tries to address important problems,
the solution seems very far from practical application to large code bases
and even attempts to translate the same language from one library to
another.
--
  Best regards
     Michal

On Sat, Jun 13, 2020 at 2:03 PM <haskell-cafe-request at haskell.org> wrote:

> Send Haskell-Cafe mailing list submissions to
>         haskell-cafe at haskell.org
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
> or, via email, send a message with subject or body 'help' to
>         haskell-cafe-request at haskell.org
>
> You can reach the person managing the list at
>         haskell-cafe-owner at haskell.org
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Haskell-Cafe digest..."
> Today's Topics:
>
>    1. transcoding - Haskell?! (Gregory Guthrie)
>    2. Re: transcoding - Haskell?! (Branimir Maksimovic)
>    3. Re: transcoding - Haskell?! (MigMit)
>    4. Re: transcoding - Haskell?! (Branimir Maksimovic)
>    5. Re: transcoding - Haskell?! (Henning Thielemann)
>
>
>
> ---------- Forwarded message ----------
> From: Gregory Guthrie <guthrie at miu.edu>
> To: "haskell-cafe at haskell.org" <haskell-cafe at haskell.org>
> Cc:
> Bcc:
> Date: Fri, 12 Jun 2020 17:33:47 +0000
> Subject: [Haskell-cafe] transcoding - Haskell?!
>
> I think it would choke on Haskell code!
>
> Haskell has so many language extensions and pragmas, and people use many
> local extensions with monads to basically create DIY DSL’s – that the code
> becomes very dense and context specific.
>
>
>
> *Facebook's TransCoder AI Converts Code From One Programming Language Into
> Another*
> *VentureBeat*
>
> * Kyle Wiggers June 9, 2020*
>
> Facebook says its TransCoder can convert code from one high-level
> programming language into another. The system, which Facebook researchers
> describe as “a neural transcompiler,” uses an unsupervised learning
> approach to translate between languages like C++, Java, and Python. The
> researchers trained TransCoder on a public GitHub corpus featuring more
> than 2.8 million open source repositories. To evaluate its capabilities,
> the researchers extracted 852 parallel functions in C++, Java, and Python
> from the online GeeksforGeeks platform and developed a new computational
> accuracy metric that tests whether hypothesis functions generate the same
> outputs as a reference when given the same inputs. Wrote the researchers,
> “TransCoder can easily be generalized to any programming language, does not
> require any expert knowledge, and outperforms commercial solutions by a
> large margin.”
>
>
>
>
>
>
>
> Dr. Gregory Guthrie
>
> Maharishi International University
>
> ----------------------------------------------------------------
>
>
>
>
>
> ---------- Forwarded message ----------
> From: Branimir Maksimovic <branimir.maksimovic at gmail.com>
> To: haskell-cafe at haskell.org
> Cc:
> Bcc:
> Date: Fri, 12 Jun 2020 19:44:01 +0200
> Subject: Re: [Haskell-cafe] transcoding - Haskell?!
>
> It would choke on assembler as well. Digital computer hard AI is
> impossible, because there is no
>
> algorithm for making algorithms. Mathematicaly impossible...
>
> Greets, Branimir.
>
>
> On 6/12/20 7:33 PM, Gregory Guthrie wrote:
>
> I think it would choke on Haskell code!
>
> Haskell has so many language extensions and pragmas, and people use many
> local extensions with monads to basically create DIY DSL’s – that the code
> becomes very dense and context specific.
>
>
>
> *Facebook's TransCoder AI Converts Code From One Programming Language Into
> Another*
> *VentureBeat*
>
> * Kyle Wiggers June 9, 2020*
>
> Facebook says its TransCoder can convert code from one high-level
> programming language into another. The system, which Facebook researchers
> describe as “a neural transcompiler,” uses an unsupervised learning
> approach to translate between languages like C++, Java, and Python. The
> researchers trained TransCoder on a public GitHub corpus featuring more
> than 2.8 million open source repositories. To evaluate its capabilities,
> the researchers extracted 852 parallel functions in C++, Java, and Python
> from the online GeeksforGeeks platform and developed a new computational
> accuracy metric that tests whether hypothesis functions generate the same
> outputs as a reference when given the same inputs. Wrote the researchers,
> “TransCoder can easily be generalized to any programming language, does not
> require any expert knowledge, and outperforms commercial solutions by a
> large margin.”
>
>
>
>
>
>
>
> Dr. Gregory Guthrie
>
> Maharishi International University
>
> ----------------------------------------------------------------
>
>
>
> _______________________________________________
> Haskell-Cafe mailing list
> To (un)subscribe, modify options or view archives go to:http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
> Only members subscribed via the mailman list are allowed to post.
>
>
>
>
> ---------- Forwarded message ----------
> From: MigMit <migmit at gmail.com>
> To: Branimir Maksimovic <branimir.maksimovic at gmail.com>
> Cc: haskell-cafe <haskell-cafe at haskell.org>
> Bcc:
> Date: Fri, 12 Jun 2020 20:14:18 +0200
> Subject: Re: [Haskell-cafe] transcoding - Haskell?!
> I wonder if this would be true:
> https://www.theolognion.com/circular-ts-js-ts-js-transpiling-yields-impressive-results/
>
> > On 12 Jun 2020, at 19:44, Branimir Maksimovic <
> branimir.maksimovic at gmail.com> wrote:
> >
> > It would choke on assembler as well. Digital computer hard AI is
> impossible, because there is no
> >
> > algorithm for making algorithms. Mathematicaly impossible...
> >
> > Greets, Branimir.
> >
> >
> >
> > On 6/12/20 7:33 PM, Gregory Guthrie wrote:
> >> I think it would choke on Haskell code!
> >> Haskell has so many language extensions and pragmas, and people use
> many local extensions with monads to basically create DIY DSL’s – that the
> code becomes very dense and context specific.
> >>
> >> Facebook's TransCoder AI Converts Code From One Programming Language
> Into Another
> >> VentureBeat
> >> Kyle Wiggers
> >> June 9, 2020
> >>
> >> Facebook says its TransCoder can convert code from one high-level
> programming language into another. The system, which Facebook researchers
> describe as “a neural transcompiler,” uses an unsupervised learning
> approach to translate between languages like C++, Java, and Python. The
> researchers trained TransCoder on a public GitHub corpus featuring more
> than 2.8 million open source repositories. To evaluate its capabilities,
> the researchers extracted 852 parallel functions in C++, Java, and Python
> from the online GeeksforGeeks platform and developed a new computational
> accuracy metric that tests whether hypothesis functions generate the same
> outputs as a reference when given the same inputs. Wrote the researchers,
> “TransCoder can easily be generalized to any programming language, does not
> require any expert knowledge, and outperforms commercial solutions by a
> large margin.”
> >>
> >>
> >>
> >>
> >> Dr. Gregory Guthrie
> >> Maharishi International University
> >> ----------------------------------------------------------------
> >>
> >>
> >>
> >> _______________________________________________
> >> Haskell-Cafe mailing list
> >> To (un)subscribe, modify options or view archives go to:
> >>
> >> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
> >>
> >> Only members subscribed via the mailman list are allowed to post.
> >>
> > _______________________________________________
> > Haskell-Cafe mailing list
> > To (un)subscribe, modify options or view archives go to:
> > http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
> > Only members subscribed via the mailman list are allowed to post.
>
>
>
>
>
> ---------- Forwarded message ----------
> From: Branimir Maksimovic <branimir.maksimovic at gmail.com>
> To: MigMit <migmit at gmail.com>
> Cc: haskell-cafe <haskell-cafe at haskell.org>
> Bcc:
> Date: Fri, 12 Jun 2020 20:25:58 +0200
> Subject: Re: [Haskell-cafe] transcoding - Haskell?!
> As long as there is math behind that it could work. There is no
> algorithm for creative thinking. Same way that if you put it on rng 1000
> monkeys will sooner or later create Shakespeare :P
>
> Greets, Branimir.
>
> On 6/12/20 8:14 PM, MigMit wrote:
> > I wonder if this would be true:
> https://www.theolognion.com/circular-ts-js-ts-js-transpiling-yields-impressive-results/
> >
> >> On 12 Jun 2020, at 19:44, Branimir Maksimovic <
> branimir.maksimovic at gmail.com> wrote:
> >>
> >> It would choke on assembler as well. Digital computer hard AI is
> impossible, because there is no
> >>
> >> algorithm for making algorithms. Mathematicaly impossible...
> >>
> >> Greets, Branimir.
> >>
> >>
> >>
> >> On 6/12/20 7:33 PM, Gregory Guthrie wrote:
> >>> I think it would choke on Haskell code!
> >>> Haskell has so many language extensions and pragmas, and people use
> many local extensions with monads to basically create DIY DSL’s – that the
> code becomes very dense and context specific.
> >>>
> >>> Facebook's TransCoder AI Converts Code From One Programming Language
> Into Another
> >>> VentureBeat
> >>> Kyle Wiggers
> >>> June 9, 2020
> >>>
> >>> Facebook says its TransCoder can convert code from one high-level
> programming language into another. The system, which Facebook researchers
> describe as “a neural transcompiler,” uses an unsupervised learning
> approach to translate between languages like C++, Java, and Python. The
> researchers trained TransCoder on a public GitHub corpus featuring more
> than 2.8 million open source repositories. To evaluate its capabilities,
> the researchers extracted 852 parallel functions in C++, Java, and Python
> from the online GeeksforGeeks platform and developed a new computational
> accuracy metric that tests whether hypothesis functions generate the same
> outputs as a reference when given the same inputs. Wrote the researchers,
> “TransCoder can easily be generalized to any programming language, does not
> require any expert knowledge, and outperforms commercial solutions by a
> large margin.”
> >>>
> >>>
> >>>
> >>>
> >>> Dr. Gregory Guthrie
> >>> Maharishi International University
> >>> ----------------------------------------------------------------
> >>>
> >>>
> >>>
> >>> _______________________________________________
> >>> Haskell-Cafe mailing list
> >>> To (un)subscribe, modify options or view archives go to:
> >>>
> >>> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
> >>>
> >>> Only members subscribed via the mailman list are allowed to post.
> >>>
> >> _______________________________________________
> >> Haskell-Cafe mailing list
> >> To (un)subscribe, modify options or view archives go to:
> >> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
> >> Only members subscribed via the mailman list are allowed to post.
>
>
>
>
> ---------- Forwarded message ----------
> From: Henning Thielemann <lemming at henning-thielemann.de>
> To: Gregory Guthrie <guthrie at miu.edu>
> Cc: haskell-cafe at haskell.org
> Bcc:
> Date: Fri, 12 Jun 2020 21:51:30 +0200 (CEST)
> Subject: Re: [Haskell-cafe] transcoding - Haskell?!
>
> On Fri, 12 Jun 2020, Gregory Guthrie wrote:
>
> > I think it would choke on Haskell code!
> >
> > Haskell has so many language extensions and pragmas, and people use many
> > local extensions with monads to basically create DIY DSL’s – that the
> > code becomes very dense and context specific.
>
> It could have its use to convert legacy code in legacy languages to shiny
> new Haskell code. :-)_______________________________________________
> Haskell-Cafe mailing list
> Haskell-Cafe at haskell.org
> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
>


-- 
  Pozdrawiam
    Michał
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/haskell-cafe/attachments/20200613/1408d88d/attachment.html>


More information about the Haskell-Cafe mailing list