Logs for the Monero Research Lab Meeting Held on 2019-07-08

作者: el00ruobuob / sarang

Logs

<sarang> Righto, let's begin our meeting!
<sarang> Agenda: https://github.com/monero-project/meta/issues/368
<sarang> Starting now with GREETINGS
<sarang> hi
<sgp_> Hi
<sarang> suraeNoether: you here?
<sarang> It's quiet… too quiet…
<suraeNoether> yes
<suraeNoether> sorry
<suraeNoether> hellow gents
<suraeNoether> harrow
<sarang> Let's jump into ROUNDTABLE then, with our small crowd
<suraeNoether> mine is going to be super fast: last week basically right after the research meeting i started getting very ill. long story short, i went to the hospital, and let me just say: recreational pancreatitis is not a thing for a reason
<sarang> but it'd be a cool name for a band
<suraeNoether> yes. yes. *strokes chin*
<suraeNoether> so, i'm trying to take it easy and i did very little this past week
<sarang> Glad to see you're well enough to debug :)
<sarang> For me, MLSAG/CLSAG verification updates continue, with PR 5707 open for review and similar changes to my CLSAG branch (to be included in later code for review)
<suraeNoether> i have a few things on my plate for today, but other than that: i'm basically all action items and no progress compared to last week. onto sarang :D
<sarang> I've been in contact with Aram, the author of the Lelantus paper/protocol; he came up with an interesting idea to make the prover very efficient, at the cost of proof size and verification
<sarang> although the verification cost can be batched, of course
<sarang> I view proving time as generally unimportant (to an extent), but it's a very clever new way to prove a 1-of-N zero commitment
<suraeNoether> good prover times means fast construction of transaction for mobile devices
<suraeNoether> which is v nice
<sarang> true
<sarang> but if it costs you both space and verification time…
<suraeNoether> unlike, say, verification time, which puts a constraint on how rapidly the network can grow, which has security consequences for the chain
<suraeNoether> ^ ah yeah that's true: is it faster with a big space tradeoff?
<sarang> Faster prover, slower verifier, bigger proof
<sarang> I think you can batch away some of the verification increase (you effectively do two smaller proofs)
<sarang> There's a non-public draft writeup already, but I assume he'll work it into the main paper once the modified security proofs are complete
<suraeNoether> oooof
<suraeNoether> very interesting consequences
<sarang> Regardless, it's a damn clever construction
<suraeNoether> i'm excited to read all about it
<sarang> I'll ask if I can send it to you suraeNoether (not public though, sorry)
<sarang> I'm investigating a possible modification to Omniring that splits out the range proofs, improving verification batching at the cost of proof size
<sarang> And, because of the Omniring non-batching currently available, am revisiting analysis of RCT3
<sarang> Which, while it would require a separate output pool (non-compatible key image structure), does allow for batching of proofs (aside from ring member group elements, which cannot be batched unless reused)
<suraeNoether> last we spoke about this, we were still interested in writing up a comparison paper, but you've done all the legwork on it so far
<suraeNoether> still the plan?
<sarang> I don't consider a formal paper necessary, or even a great use of time
<sarang> But analyses of spacetime, totally
<suraeNoether> fair, maybe we can do a blog post on tradeoffs between the three schemes or something
<sarang> Maybe, but it gets subtle and complex really quickly under many different assumptions
<sarang> There isn't really a quick-and-dirty soundbite answer to which is better or worse
<sarang> Depends heavily on input/output structure, use of fixed epochs, batch behavior, etc.
<sarang> and output pool migration is nontrivial
<sarang> Omniring would _not_ require this… Lelantus and RCT3 would
<sarang> (or rather, Omniring does not _require_ this)
<suraeNoether> see… when you say all that, it seems like it *is* a good use of time. maybe not high priority, but
<sarang> A comparison is useful, I agree. But I don't want it to get lost in unnecessary formality of a full paper
<sarang> And a comparison is exactly what I'm doing
<sarang> Any questions on these topics?
<sgp_> Just a comment to say simple comparisons are good
<sarang> roger
<sarang> Does anyone else have research topics of interest to share?
<sarang> Righto!
<nioc> suraeNoether: people are wondering what the attendance was at konferenco
<sarang> At least provide a range proof
<moneromooo> 26 people and 172 sybils.
<nioc> an article in coindesk mentioned 75 which I know is way low
<suraeNoether> nioc: i'm finishing up my post-mortem report on the konferenco today (on my action item list). We had 150 swag bags made, with around 30-40 leftover, but we had 27 speakers and like 10 sponsors.
<nioc> I estimated 120 just by glancing
<suraeNoether> 75 is the number i gave coindesk for the number of attendees on Saturday morning, but more people bought tickets on both days and the totals were higher
<suraeNoether> if you count speakers and sponsors, that's around 110 on the first day, and around 125 the second
<nioc> think you mran Sunday morning
<nioc> mean
<suraeNoether> nope, we sold tickets throughout the afternoon on saturday and a few on sunday too. but coindesk asked for a comment on saturday morning, so i told them what i had sold at that point
<nioc> I'll wait for the report
<suraeNoether> k
<nioc> thx
<sarang> suraeNoether: congratulations on effectively committing yourself to running a kickass conference annually until the end of time =p
<suraeNoether> it was actually some of the best days of my life, but i've been told explicitly by my doctors that i need to take a vacation
<suraeNoether> so i'm planning for that in august, since scary cardio and internal medicine people told me so with stern voices
<sarang> Well then, let's move on to ACTION ITEMS
<sarang> I have many things in progress. Lelantus proof review, modified Omniring split proof analysis, RCT3 analysis, and starting to put together my defcon talk/workshop
<suraeNoether> my action items: konferenco post mortem, research report for previous quarter, funding request for the next 3 months, and some debuggin
<sarang> I'm doing a talk on transaction protocols (very high level), a workshop on simple cryptographic constructions with Python, and a panel discussion at the blockchain village
<suraeNoether> oh and i'm definitely not going to defcon this year. i can ship leftover swag like our USB data blockers with the monero logo and our pull-up banners if someone sends me the information for it
<sarang> :(
<sarang> The pull-up banners would be nice, assuming it's cheaper to ship than to get new ones
<sarang> as would the USB blockers
<suraeNoether> i'll look into it; ordering the usb blockers may be short notice but i can find out
<suraeNoether> unless you just meant shipping the banners
<suraeNoether> which is cool too
<sarang> Yeah I meant the banners
<sarang> I believe there was an idea to perhaps order more blockers for this (I'm not the one to ask)
<sarang> If you have extra USB blockers and would ship with the banners, cool
<suraeNoether> fantastic, i am happy folks liked the blockers
<sarang> Any last questions or comments before we formally adjourn, since agenda topics have been completed?
<sarang> Going once
<sarang> twice
<sarang> adjourned!
<sarang> Thanks to everyone for attending; logs will be posted shortly on the github issue


Post tags : Dev Diaries, Cryptography, Monero Research Lab