Monday, November 08, 2021

Postdoc call for FODSI

As a member of FODSI (Foundations of Data Science Institute -- an NSF funded institute with the aim of advancing theoretical foundations for data science), I'm self-interestedly posting the call for postdocs for this year.  Two of the areas are  a) Sketching, Sampling, and Sublinear-Time Algorithms   and b)  Machine Learning for Algorithms (which includes what I call "Algorithms with Predictions.")  I'd be happy to see postdoc applications in those areas from people who want to spend some time at Harvard, for example.... but of course there are lots of other exciting things going on with FODSI too and you should take a look.

The call is at https://academicjobsonline.org/ajo/jobs/20132  

Call text below:

The Foundations of Data Science Institute (FODSI), funded by the National Science Foundation TRIPODS program, is announcing a competitive postdoctoral fellowship. FODSI is a collaboration between UC Berkeley and MIT, partnering with Boston University, Northeastern University, Harvard University, Howard University and Bryn Mawr College. It provides a structured environment for exploring interdisciplinary research in foundations of Data Science spanning Mathematics, Statistics, Theoretical Computer Science and other fields.

We are looking for multiple postdoctoral team members who will collaborate with FODSI researchers at one or more of the participating institutions. These positions emphasize strong mentorship, flexibility, and breadth of collaboration opportunities with other team members -- senior and junior faculty, postdocs, and graduate students at various nodes around the country. Furthermore, postdoctoral fellows will be able to participate in workshops and other activities organized by FODSI.

The fellowship is a one-year full-time appointment, with the possibility of renewal for a second year (based upon mutual agreement) either at the same or at another FODSI institution. The start date is flexible, although most appointments are expected to start in summer 2022. Candidates are encouraged to apply to work with more than one faculty mentor at one or more participating institutions (in-person mentoring is preferred, but remote options will be also considered). The applicants should have an excellent theoretical background and a doctorate in a related field, including Mathematics, Statistics, Computer Science, Electrical Engineering or Economics. We particularly encourage applications from women and minority candidates.

The review process will start on November 15, 2021 and will continue until positions are filled.

Thursday, November 04, 2021

HotNets Presentation : Zero-CPU Collection with Direct Telemetry Access

HotNets has asked that we let people know that the 2021 presentations are available here.  I'm using that an excuse to highlight our paper on Zero-CPU Collection with Direct Telemetry Access (arxiv version here), but really I want to highlight the talk by graduate student Jonatan Langlet (Queen Mary University of London) who, as is the nature of graduate students, did all of the real work, and who really did a great job on the talk (direct link).  If you guessed from my involvement this involves hashing in some way, your maximum likelihood estimate turns out to be correct.

I think our work fits the HotNets call, which asks for new approaches and preliminary work.  Specifically, the call for the HotNets workshop says this:

We invite researchers and practitioners to submit short position papers. We encourage papers that identify fundamental open questions, advocate a new approach, offer a constructive critique of the state of networking research, re-frame or debunk existing work, report unexpected early results from a deployment, report on promising but unproven ideas, or propose new evaluation methods. Novel ideas need not be supported by full evaluations; well-reasoned arguments or preliminary evaluations can support the possibility of the paper’s claims.

We seek early-stage work, where the authors can benefit from community feedback. An ideal submission has the potential to open a line of inquiry for the community that results in multiple conference papers in related venues (SIGCOMM, NSDI, CoNEXT, SOSP, OSDI, MobiCom, MobiSys, etc.), rather than a single follow-on conference paper. The program committee will explicitly favor early work and papers likely to stimulate reflection and discussion over “conference papers in miniature”.

There are similar other "Hot" workshops in other areas, and it was about 14 years ago that I asked whether CS theory should have a HotTheory workshop.  There's been a proliferation of new conferences and workshops in theory since then, but none of them really seem to have this flavor.  So maybe it's worth asking again whether a HotTheory workshop would make sense?  Or do existing theory events meet the theory community needs?