As a mentor I would say it is certainly possible to outperform exsting mega-solutions in some narrow domain.<br>Just as I did with hPDB <a href="https://hackage.haskell.org/package/hPDB">https://hackage.haskell.org/package/hPDB</a><br>But it requires a lot of skill and patiece.<br><br>Please proceed with this project with the current list of mentors.<br>I think me and Dominic have already declared committment.<br><br>You might also start by making a table of best competing solutions in other. languages, their respective strengths, and ways that we can possibly improve on them!<br><br>Where do You keep Your application draft? Ideally it should be a shared space where you can add mentors as co-editors.<br>—<br> Cheers<br> Michal<br><div class="gmail_quote"><div dir="ltr">On Sun, 25 Mar 2018 at 02:32, <<a href="mailto:dominic@steinitz.org">dominic@steinitz.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div style="word-wrap:break-word;line-break:after-white-space">The list of mentors for this project looks great to me. I am not sure if I can add much other than I think this is a nice project. Perhaps it would be best to get the advice of some of the mentors?<div><br></div><div>For some very simple tests with an ODE solver, I concluded that accelerate can perform at least as well as Julia. It would certainly be very helpful to be able to get Jacobians for ODE solving and for other applications.</div><div><br><div>
<div style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px">Dominic Steinitz</div><div style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px"><a href="mailto:dominic@steinitz.org" target="_blank">dominic@steinitz.org</a></div><div style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px"><a href="http://idontgetoutmuch.wordpress.com" target="_blank">http://idontgetoutmuch.wordpress.com</a></div><div style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px">Twitter: @idontgetoutmuch</div><div style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px"><br></div><br class="m_3152147633407713228Apple-interchange-newline">
</div>
<div><br><blockquote type="cite"><div>On 24 Mar 2018, at 17:20, Charles Blake <<a href="mailto:cb307@st-andrews.ac.uk" target="_blank">cb307@st-andrews.ac.uk</a>> wrote:</div><br class="m_3152147633407713228Apple-interchange-newline"><div><div id="m_3152147633407713228divtagdefaultwrapper" dir="ltr" style="font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;font-size:12pt;font-family:Calibri,Helvetica,sans-serif"><p style="margin-top:0px;margin-bottom:0px"></p><div>Thanks for the response Michal,</div><div><br></div><div>Yes, this did cross my mind - and I wouldn't be expecting to outperform those frameworks in the timeframe available! I assumed that the reason that this project was suggested was perhaps:</div><div><br></div><div>a) there is some intrinsic value in implementing these algorithms natively in haskell (hence why the 'ad' library was developed in the first place), so that those who want to use parallel automatic differentiation / the machine learning algorithms built on top of it can do so without leaving the haskell ecosystem,</div><div><br></div><div>and b) because the challenges involved in implementing parallel ad in a purely functional language are a little different to those involved in doing so in OO/imperative languages - so it might be interesting from that angle as well?</div><div><br></div><div>So perhaps my aim would no be to do something unique, but rather to do something that has already done well in other languages, but has not yet been provided as a haskell library. Does this sound like a reasonable approach or do I need to find a slightly more unique angle?</div><div><br></div><div>Thanks,</div><div>Charlie</div><p style="margin-top:0px;margin-bottom:0px"></p></div><hr style="font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;display:inline-block;width:617.390625px"><span style="font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;float:none;display:inline!important"></span><div id="m_3152147633407713228divRplyFwdMsg" dir="ltr" style="font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px"><font face="Calibri, sans-serif" style="font-size:11pt"><b>From:</b><span class="m_3152147633407713228Apple-converted-space"> </span>Michal J Gajda <<a href="mailto:mgajda@mimuw.edu.pl" target="_blank">mgajda@mimuw.edu.pl</a>><br><b>Sent:</b><span class="m_3152147633407713228Apple-converted-space"> </span>24 March 2018 16:56:35<br><b>To:</b><span class="m_3152147633407713228Apple-converted-space"> </span>Dominic Steinitz; Marco Zocca; <a href="mailto:accelerate-haskell@googlegroups.com" target="_blank">accelerate-haskell@googlegroups.com</a>; Charles Blake; <a href="mailto:haskell-cafe@haskell.org" target="_blank">haskell-cafe@haskell.org</a><br><b>Subject:</b><span class="m_3152147633407713228Apple-converted-space"> </span>Re: Accelerating Automatic Differentiation</font><div> </div></div></div></blockquote></div></div></div><div style="word-wrap:break-word;line-break:after-white-space"><div><div><blockquote type="cite"><div><div style="font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px"><div><div dir="auto">Hi Charlie</div><div dir="auto"><br></div><div dir="auto">It certainly looks like exciting project, but the bar is currently placed very high.</div><div dir="auto">TensorFlow package not only provides automatic differentiation for whole programs, but also optimizes data processing both on GPU, and reading to achieve large batches.</div><div dir="auto">This field has a lot of hot developments, so You would either need to propose something unique to Haskell, or You risk being outclassed by PyTorch and TensorFlow bindings</div><div dir="auto"> Maybe Dominic suggests something too.</div><div dir="auto"> Cheers</div></div><div><div dir="auto"> Michal</div></div></div></div></blockquote></div></div></div></blockquote></div>