Re: [pl-seminar] Scheduling PL-Seminar for Spring Semester

[Alisa Maas <ajmaas@xxxxxxxxxxx>]

Tom Reps will be giving the first talk. The abstract is below. The sign up sheet for talks is here: https://docs.google.com/spreadsheets/d/1XO4sU5IXEra5QoWqMyW_Jzg9qi5fm4S3jPsRjIjwnDw/edit#gid=0, and Aws has set up a spreadsheet for tracking talks from POPL â16 that we might want to hear more about (https://docs.google.com/spreadsheets/d/1IcZojAe0yT6QAyj_HpXHbapNVdEQBguNfoTe6xw_Q1w/edit?usp=sharing). If youâre looking for a topic, consider one of those. 

We will meet at noon in CS 4310.

 Newtonian Program Analysis via Tensor Product

                          Thomas Reps
          University of Wisconsin and GrammaTech, Inc.

Recently, Esparza et al. generalized Newton's method -- a
numerical-analysis algorithm for finding roots of real-valued
functions -- to a method for finding fixed-points of systems of
equations over semirings.  Their method provides a new way to solve
interprocedural dataflow-analysis problems.  As in its real-valued
counterpart, each iteration of their method solves a simpler
``linearized'' problem.

One of the reasons this advance is exciting is that some numerical
analysts have claimed that ```all' effective and fast iterative
[numerical] methods are forms (perhaps very disguised) of Newton's
method.''  However, there is an important difference between the
dataflow-analysis and numerical-analysis contexts: when Newton's
method is used on numerical-analysis problems, multiplicative
commutativity is relied on to rearrange expressions of the form ``c*X
+ X*d'' into ``(c+d) * X.''  Such equations correspond to path
problems described by regular languages.  In contrast, when Newton's
method is used for interprocedural dataflow analysis, the
``multiplication'' operation involves function composition, and hence
is non-commutative: ``c*X + X*d'' cannot be rearranged into ``(c+d) *
X.''  Such equations correspond to path problems described by linear
context-free languages (LCFLs).

In this talk, we present an improved technique for solving the LCFL
sub-problems produced during successive rounds of Newton's method. The
method applies to predicate abstraction, on which most of today's
software model checkers rely.

(Joint work with Emma Turetsky and Prathmesh Prabhu.)

On Feb 3, 2016, at 3:02 PM, Loris D'Antoni <loris@xxxxxxxxxxx> wrote:

Let's do 4310.

Thanks,

-Loris

On Wed, Feb 3, 2016 at 2:45 PM, Drew Davidson <drew.j.davidson@xxxxxxxxx> wrote:

Both 4310 (for a larger room) and 5331 (for a cozier room) are open in that time slot for every week I checked. 

I greedily grabbed both, but we should release the one that we don't need.

On Wed, Feb 3, 2016 at 2:05 PM, Loris D'Antoni <loris@xxxxxxxxxxx> wrote:

Tue 12-1 seems like a winner. Only exception, Stephen cannot make it until 12.30. We'll talk about this during the first meeting and see if we can delay a bit.

Why don't we start next Tuesday?

Do we have a room?

-Loris

On Mon, Feb 1, 2016 at 12:07 AM, Alisa Maas <ajmaas@xxxxxxxxxxx> wrote:

Tom Reps has offered already to give the first talk on "Newtonian Program Analysis via Tensor Productâ, from POPL 2016. The only times that work for the majority of people seem to be Tuesdays or Wednesdays at 5pm. If you have a preference between these two times and have not yet filled out the form, please do so asap. 

On Jan 31, 2016, at 8:35 PM, Aws Albarghouthi <aws@xxxxxxxxxxx> wrote:

Thanks, Alisa, for restarting this. I was thinking we could for the first few meetings read papers from POPL 2016 that we thought were interesting. I started a Google Doc where I listed the papers that I myself would like to read soon:

https://docs.google.com/spreadsheets/d/1IcZojAe0yT6QAyj_HpXHbapNVdEQBguNfoTe6xw_Q1w/edit?usp=sharing

If interested in doing this, please enter the papers you'd like to see in the seminar or add 1 to the interest column of existing papers.

On Fri, Jan 29, 2016 at 1:17 PM, Alisa Maas <ajmaas@xxxxxxxxxxx> wrote:

Hi all,

Iâve been asked to schedule PL seminar this semester. Please select the times that work for you, keeping in mind that each slot is an hour long. http://whenisgood.net/plseminar. 

After we have a time slot selected, Iâll send out the link so people can sign up to present papers. You donât have to be an author on the paper to present!

One more thing: to avoid missing anyone on the email list, make sure youâre all on the pl-seminar list, not just pl-reading. Iâm sending to both this time, but it would be nice to have just one list in the future, and I donât believe these are used for different purposes normally. 

Thanks,

Alisa

_______________________________________________
Pl-seminar mailing list
Pl-seminar@xxxxxxxxxxx
https://lists.cs.wisc.edu/mailman/listinfo/pl-seminar

_______________________________________________
Pl-seminar mailing list
Pl-seminar@xxxxxxxxxxx
https://lists.cs.wisc.edu/mailman/listinfo/pl-seminar

_______________________________________________
Pl-seminar mailing list
Pl-seminar@xxxxxxxxxxx
https://lists.cs.wisc.edu/mailman/listinfo/pl-seminar