<div><div dir="auto">just fyi.</div></div><div dir="auto"><br></div><div dir="auto">i believe the infrastructure task force now has the authority to accept associate partners.</div><div dir="auto"><br></div><div dir="auto">and regarding mark, i think we have previously decided to welcome his association, so i will just go ahead and ask him to request accounts.</div><div dir="auto"><br></div><div dir="auto">like with robert, i will encourage them to put load on Abel first. once the request an allocation on Taito, i imagine we can deal with that separately :-).</div><div dir="auto"><br></div><div dir="auto">best wishes, oe</div><div dir="auto"><br></div><div dir="auto">ps: i am eager to get mark associated, in part because i suspect it strengthens our case in asking NeIC for continued support.</div><div dir="auto"><br></div><div><br><div class="gmail_quote"><div dir="ltr">---------- Forwarded message ---------<br>From: <strong class="gmail_sendername" dir="auto">Mark Fishel</strong> <span dir="ltr"><<a href="mailto:fishel@ut.ee">fishel@ut.ee</a>></span><br>Date: Sun, 21 Oct 2018 at 06:44<br>Subject: Re: [NLPL Board] NLP in Tartu, Estonia<br>To: Stephan Oepen <<a href="mailto:oe@ifi.uio.no">oe@ifi.uio.no</a>><br></div><br><br><div dir="ltr"><span style="font-size:12.8px">Stephan,</span><div><span style="font-size:12.8px"><br></span></div><div><span style="font-size:12.8px">I forgot to include some further details, described on the NLPL website.</span></div><div><span style="font-size:12.8px"><br></span></div><div><span style="font-size:12.8px">1. expected types of computing</span></div><div><span style="font-size:12.8px">Training and evaluating neural models for translation, text-to-speech.</span></div><div><span style="font-size:12.8px"><br></span></div><div><span style="font-size:12.8px">2. software</span></div><div><span style="font-size:12.8px">We use and write open source software. More specifically,</span></div><div><span style="font-size:12.8px">- for NMT we use SockEye from AWSLabs</span>, which uses mxnet as the low-level backend</div><div>- for TTS and for all custom models we use either PyTorch or Tensorflow</div><div>- our own software is at <a href="https://github.com/TartuNLP" target="_blank">https://github.com/TartuNLP</a> </div><div><br></div><div><span style="font-size:12.8px">3. data</span></div><div><span style="font-size:12.8px">This includes several different datasets. For machine translation we use OPUS's corpora a lot (Europarl, OpenSubtitles, others). Our group has annotated and developed a large number of Estonian corpora</span>, a lot of which can be accessed through <a href="https://metashare.ut.ee/" target="_blank">https://metashare.ut.ee/</a>, maintained by the Center of Estonian Language Resources.</div><div><span style="font-size:12.8px"><br></span></div><div><span style="font-size:12.8px">4. anticipated group of users</span></div><div>The affiliation for all the users from our side will be the NLP research group, institute of computer science, University of Tartu. The number of active users at the moment would be 8 people, which is bound to grow in th enear future, and of course we can scale this down if needed.</div><div><br></div><div>Best wishes,<br>Mark</div></div><div dir="ltr"><div><br><br><div class="gmail_quote"><div dir="ltr">On Sun, Oct 21, 2018 at 7:32 AM Mark Fishel <<a href="mailto:fishel@ut.ee" target="_blank">fishel@ut.ee</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">Dear Stephan, everyone,<div><br></div><div>we are definitely interested! Sorry for the huge delay in my answer.</div><div><br></div><div>Our primary need is computational power, mainly servers with GPUs. We have an HPC at our university, but its GPU capacity is only enough for small-scale experiments and several groups use it besides us.</div><div><br></div><div>The experiments where we would need this computational power is currently neural machine translation and end-to-end speech synthesis, I am hoping to expand it in the near future to dependency parsing and other topics.</div><div><br></div><div>In addition to the computational resources another big appeal to me is the possibility of more easily exchanging data, experimental setup and results with partners from the NLPL network, we will gladly expand the collaboration with colleagues from Helsinki University and Uppsala, and of course will happily start new collaborations with other partners.</div><div><br></div><div>Please let me know if anything else is needed from my side, I promise no more delays :-)</div><div><br></div><div>Best wishes,<br>Mark</div></div><br><div class="gmail_quote"><div dir="ltr">On Tue, Oct 9, 2018 at 12:10 PM Stephan Oepen <<a href="mailto:oe@ifi.uio.no" target="_blank">oe@ifi.uio.no</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div>hi again, mark,</div><div><br></div><div>re-sending the message below, to make sure it actually made it to you? i took it as all but a certainty that you were interested in joining the NLPL associate programme, but to initiate that process it would be good to have a brief summary of your needs, according to the guidelines for associate partners on the NLPL web site.</div><div><br></div><div>best wishes, oe<br></div></div><div dir="ltr"><div></div><div><br></div><div><br><div class="gmail_quote"><div dir="ltr">---------- Forwarded message ---------<br>From: <b class="gmail_sendername" dir="auto">Stephan Oepen</b> <span dir="ltr"><<a href="mailto:oe@ifi.uio.no" target="_blank">oe@ifi.uio.no</a>></span><br>Date: Tue, Sep 11, 2018 at 5:36 PM<br>Subject: Re: [NLPL Board] NLP in Tartu, Estonia<br>To: Mark Fishel <<a href="mailto:fishel@ut.ee" target="_blank">fishel@ut.ee</a>><br>Cc: Bjørn Lindi <<a href="mailto:bjorn.lindi@ntnu.no" target="_blank">bjorn.lindi@ntnu.no</a>>, <a href="mailto:contact@nlpl.eu" target="_blank">contact@nlpl.eu</a> <<a href="mailto:contact@nlpl.eu" target="_blank">contact@nlpl.eu</a>><br></div><br><br>hi again, mark,<br> <br> as of this summer, NLPL has created an associate programme, to give<br> additional compute-intensive NLP research groups in northern europe a<br> way of taking advantage of our software and data installations. the<br> notion of associate partners was minted in response to your original<br> query in april, so i hope you might still be interested? if so,<br> please see the instructions on the NLPL front page<br> (‘<a href="http://www.nlpl.eu" rel="noreferrer" target="_blank">http://www.nlpl.eu</a>’) and send us an email about your anticipated<br> needs.<br> <br> best wishes, oe<br> <br> <br> On Fri, Apr 13, 2018 at 8:50 AM, Mark Fishel <<a href="mailto:fishel@ut.ee" target="_blank">fishel@ut.ee</a>> wrote:<br> > Dear Bjørn,<br> ><br> > thank you for the info! I will see about the Taito system, and if there are<br> > any updates on outside groups joining, I would be happy to find out!<br> ><br> > Best wishes,<br> > Mark<br> ><br> ><br> > On Wed, Apr 11, 2018 at 9:53 AM Bjørn Lindi <<a href="mailto:bjorn.lindi@ntnu.no" target="_blank">bjorn.lindi@ntnu.no</a>> wrote:<br> >><br> >> Dear Mark,<br> >> the NLPL have just started a discussion on how to open our resources to<br> >> research groups outside NLPL. This is something we will work on in the<br> >> coming months, though one way to get an immediate start is to get an<br> >> personal account on the Finnish system taito. A personal account comes with<br> >> a low compute quota (I think it is around 10 000 CPU hours), but you will<br> >> be able to browse NLPL resources and see what we currently provide.<br> >><br> >> Be sides this practical step, we will be able to take you into account as<br> >> we investigate how our resources could be shared to a greater benefit.<br> >><br> >> Thanks for reaching out. I am sure we will find a way to collaborate.<br> >><br> >> Yours Sincerely<br> >> Bjørn Lindi<br> >> NLPL Project Manager<br> >><br> >><br> >><br> >> On 10 Apr 2018, at 19:03, Mark Fishel <<a href="mailto:fishel@ut.ee" target="_blank">fishel@ut.ee</a>> wrote:<br> >><br> >> Dear NLPL people,<br> >><br> >> Estonia is currently looking for ways of closer ties with NEIC, and I was<br> >> wondering if it is possible for the NLP group in Tartu to collaborate with<br> >> NLPL, and what could the conditions for that be?<br> >><br> >> The NLP group in Tartu (<a href="https://nlp.cs.ut.ee/" rel="noreferrer" target="_blank">https://nlp.cs.ut.ee/</a>) is working on data science<br> >> applied to NLP, as well as linguistic resources like UD and other corpora.<br> >> In particular this year we are organizing two sub-tracks in the translation<br> >> shared task of WMT (unsupervised and multilingual NMT) and have participated<br> >> in the tasks on translation, metrics and quality estimation.<br> >><br> >> As far as infrastructure goes, we have a high performance computing center<br> >> (<a href="http://hpc.ut.ee" rel="noreferrer" target="_blank">http://hpc.ut.ee</a>), but could use much more than what we currently have.<br> >><br> >> So, I would love to talk to someone in NLPL and talk about possibilities /<br> >> requirements / etc.<br> >><br> >> Best wishes,<br> >> Mark<br> >><br> >><br> ><br> </div></div></div></blockquote></div> </blockquote></div></div></div> </div></div>