Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: What was the best CS paper you read in 2017?

news.ycombinator.com

Post reply on HN
Print view View original

111–120 of 131 posts

Re: Ask HN: What was the best CS paper you read in 2017?

#111

Post by chubot »

yazr wrote: post #43
chubot wrote: post #27

Earlier quoted context omitted.

I'm gonna throw some cold water on this and say this is not a new paradigm by any means. https://www.sigarch.org/the-unreasonable-ineffectiveness-of-... However, it's certainly possible that the time for this idea has come. Google is probably in the best position to apply it. I will say that after having worked at Google for over a decade, some of it on data center performance, there are plenty of inefficiencies that…

Can you please give some examples of resource-mis-allocations ?! AFAIK storage is not the system bottle it used to be. We always want more, but network and cores are relatively plentiful. If we could magically (and safely) modify the software stack, which areas could give x2 or x3 improvements ?

As far as Google goes, the easiest place to get better end-user latency/performance (2x-3x) is ... fixing the JavaScript.

I'm being totally serious. Backends are generally fast, and the backend engineers are performance-minded.

Front end engineers are not as cognizant of performance (somewhat necessarily, since arguably they have a harder problem to solve). Back in the mid-2000's, Gmail/Maps/Reader days Google had a lot of great JS talent, but it seems to have ceded some of that ground to Facebook and Microsoft.

If you have heard Steve Souders speak, he always mentions that he was a backend guy. Until he actually measured latency, and realized that the bottleneck is the front end. That was at Yahoo, but it's very much true for Google too.

http://stevesouders.com/bio.php

I would like to see a machine learning system rewrite JavaScript code to perform better and make UI more usable. I believe that's beyond the state of the art now, but it's probably not out of the question in the near future.

-----

As far as scheduling that was just one example of an important systems problem that hasn't been solved with machine learning. Not saying it can't be, of course. Just that this is a research direction and not a deployed system.

It's also important to note that there are plenty of other feedback-based/data-driven algorithms for resource management that are not neural nets. If neural nets work, then some simpler technique probably works too.

Re: Ask HN: What was the best CS paper you read in 2017?

#112

Post by yazr »

sulam wrote: post #79
yazr wrote: post #43

Earlier quoted context omitted.

Can you please give some examples of resource-mis-allocations ?! AFAIK storage is not the system bottle it used to be. We always want more, but network and cores are relatively plentiful. If we could magically (and safely) modify the software stack, which areas could give x2 or x3 improvements ?

Network latency right now is the biggest issue we have. If we could magically (using your term here!) get computational resources and data dramatically closer to end users, it would easily give 2 or 3x improvement. Doing this safely of course means consistently in this case, and being able to solve things like safe replication of large data sets. I dunno how to do it, but you asked and that's the biggest thing I can…

So backend-to-frontend latency?

Or are there also plenty server-to-server scenarios?

Re: Ask HN: What was the best CS paper you read in 2017?

#113

Post by chubot »

rgrieselhuber wrote: post #34
chubot wrote: post #27

Earlier quoted context omitted.

I'm gonna throw some cold water on this and say this is not a new paradigm by any means. https://www.sigarch.org/the-unreasonable-ineffectiveness-of-... However, it's certainly possible that the time for this idea has come. Google is probably in the best position to apply it. I will say that after having worked at Google for over a decade, some of it on data center performance, there are plenty of inefficiencies that…

Any resources you can recommend on the scheduling topic?

As mentioned, that was off the top of my head. Some of it is "inside baseball" at Google, but there are a bunch of published papers about cluster scheduling. This one is a good overview, and has numbers, evaluation, lessons learned, etc.:

https://research.google.com/pubs/pub43438.html

My overall point is that even if learned indexes replace traditional data structures (which is a big if), plenty of important systems problems that remain. Some that I could think of:

- Fixing front end performance as mentioned in a sibling comment.

- System administration of a billion Android phones. The state of these seems pretty sorry.

- Auditing app stores for security bugs.

- Eliminating security bugs by construction. Hell, even just eliminating XSS.

- Software that runs transparently in multiple data centers. (Spanner is a big step in that direction, but not a panacea.)

I could name like 10 more... none of these have anything to do with learned indexes. The point is that the OP is generalizing from a really specific technique to some kind of hand-wavy "machine learning writes all the code" secnario.

Re: Ask HN: What was the best CS paper you read in 2017?

#115

Post by NovaX »

posnet wrote: post #67

A General-Purpose Counting Filter: Making Every Bit Count A better alternative to bloom filters and they provide a well written implementation as well https://github.com/splatlab/cqf https://www3.cs.stonybrook.edu/~ppandey/files/p775-pandey.pd...

Interestingly this work has nearly identical algorithm to a previous paper. Good ideas are rediscovered.

Counting with TinyTable: Every Bit Counts! http://www.cs.technion.ac.il/users/wwwb/cgi-bin/tr-get.cgi/2...

Re: Ask HN: What was the best CS paper you read in 2017?

#116

Post by yazr »

yazr wrote: post #112
sulam wrote: post #79

Earlier quoted context omitted.

Network latency right now is the biggest issue we have. If we could magically (using your term here!) get computational resources and data dramatically closer to end users, it would easily give 2 or 3x improvement. Doing this safely of course means consistently in this case, and being able to solve things like safe replication of large data sets. I dunno how to do it, but you asked and that's the biggest thing I can…

So backend-to-frontend latency? Or are there also plenty server-to-server scenarios?

After re-reading your comment, are you referring to end-users inside the enterprise perimeter ?

e.g. devs using remote build system ? A local workload which access a mostly-remote db ?

Re: Ask HN: What was the best CS paper you read in 2017?

#117

Post by auggierose »

mcguire wrote: post #100
auggierose wrote: post #91

Earlier quoted context omitted.

Isabelle/HOL (the one sel4 was done in) appears in none of your two schools ... This might be because it is the best combination of interactive theorem proving with automated methods that currently exists out there.

As far as I know, Isabelle/HOL is fairly similar to Coq, perhaps with more automation. If there is something like Software Foundations for Isabelle, I'd be slobbering all over myself in excitement to hear about it. (And I'm still a little confused about how the C code of seL4 is verified against the formal specs. Is there something like Frama-C involved?)

Both are interactive theorem provers, so yeah. But the two school comment said Coq/dependent-types, and Isabelle/HOL is neither Coq, nor does it do dependent types :)

If you are looking for developments in Isabelle/HOL, there is the AFP: https://www.isa-afp.org/

There is also a recent book that might interest you: http://www.concrete-semantics.org/

Re: Ask HN: What was the best CS paper you read in 2017?

#118

Post by sova »

smsx wrote: post #108
sova wrote: post #62

Earlier quoted context omitted.

Yo! This is what I'm talkin' about! Replace directories with self-consistent neural nwots [neural blackbox thingie of weights and activation vals]. It's cool, but for every convenience we take in the direction HD space we backpeddle in processing [meaning, although the dir structure could be super condensed this way with a learned index, this approach is less flexible to cold or not-cpu-related analysis]. It's very c…

> but in the ideal computer we do no calculation, and that is something worth mentioning Can you expand on this bit? I'm not sure I understand what you mean

Absolutely! This is a statement on minimization of computational effort. When the solution has already been provided we need not do any calculation. One way to think about the quoted statement and how I meant it is to consider big O notation and how some operations take O(n) or (O lg n) steps dependent on input in order to reach a solution or complete a process or spit out a result; in this hypothetical best case our operations all take O(1). This is not really a calculation but in fact knowledge coming to light.

Since writing this comment I have text-to-speech'd the whole of the article and I listened to it one afternoon over smooth classical music and I have come to appreciate the fact that many computers will have neural network processors like tensor processors or at the very least really good GPU, making it all the more likely that such algorithms [especially used in the specific case of fault intolerant data organization and file hierarchy] will actually not require so much computational effort as traditionally believed.

Re: Ask HN: What was the best CS paper you read in 2017?

#119

Post by leecarraher »

leecarraher wrote: post #90

Dynamic Routing Between Capsules the long awaited paper from Geoffrey Hinton (although Sara Sabour is principal author, Hinton has been hinting at this for a while now) on an alternative to back-propagation https://arxiv.org/abs/1710.09829

who down-votes a paper suggestion, and is too afraid to explain why.
Post reply on HN