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 excit ..read more

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 ou ..read more

In this post, we'll look at the natural problem of how to communicate an estimate of a real value in [0,1], using just 1 bit.  The post is based on this paper (by Ran Ben-Basat of UCL and Shay Vargaftik of VMware Research and myself -- they helped also with the post) that appeared in ICALP last year.  This question is motivated by various aggregation problems;  multiple sending devices may measure a value, and wish to send the value to an aggregator who will compute something from the received values, such as the average.  In our problem, the senders have a real value ..read more

The last couple of years one aspect of research I've greatly enjoyed is getting back into networking, which is really due to my excellent (and patient) collaborators Ran Ben Basat (now at UCL, was a postdoc at Harvard) and Minlan Yu (at Harvard).  Minlan and I are working to establish a larger-scale Networking+Theory (hopefully broadening to an even larger Systems+Theory) group at Harvard, working on algorithmic problems in the context of real (or at least very real-ish) systems.  We have funding, and are looking for a postdoc, the basic description is below.  Ideally we're look ..read more

I was looking at today's Harvard Crimson, and noted that Harvard's shopping period looks ready to be vanished again.  Shopping period is that wonderful Harvard tradition where students don't preregister for classes, but instead they choose classes after the first week, after having a chance to go to a lecture or two and see how they like it.  I encourage students -- and faculty -- to push back against efforts that restrict student flexibility in choosing their classes.  While administrators hate it, I still think it's better for students to avoid strong forms of preregistra ..read more

I have had papers rejected, with comments of the form that the results seem too easy, and are at the level of a homework assignment.  Generally, I think these reviewers miss the point.  The fact that the results seem easy may be because the point isn't the derivation but the conception and framing of the problem.  I actually think that generally it's an interesting subclass of good papers that can be and are turned into homework assignments. A new-ish paper of mine, Queues with Small Advice, was recently accepted to the very new SIAM Conference on Applied and Computational ..read more

We're having an online workshop on "Machine Learning for Algorithms" on July 13-14, with a great group of speakers.  Announcement below, link at https://fodsi.us/ml4a.html, free registration (but please register in advance)! In recent years there has been increasing interest in using machine learning to improve the performance of classical algorithms in computer science, by fine-tuning their behavior to adapt to the properties of the input distribution. This "data-driven" or "learning-based" approach to algorithm design has the potential to significantly improve the efficiency of som ..read more

I wanted to give a pointer to a new preprint on bioRxiv on developing diagnostic assays for viruses, by (first author) Hayden Metsky (and others!) out of the Sabeti Lab at the Broad Institute (that I've been a bit involved with).  Hayden, who somehow is both a computer science PhD and an expert in virology, has devised a novel software pipeline for developing diagnostics that are designed from the start to deal with genomic diversity (a virus evolves to have many somewhat different variants) and the challenge of false matches (you don't want to get false positives from matching ..read more

I wanted to link to a survey that is up entitled Committee on TCS Connections Questionnaire.  They are examining modifying approaches to publishing in the theoretical computer science community, and they are focusing on FOCS/STOC. I personally approve of the idea of the committee, though I admit I am concerned that it's too little, too late.  For years, FOCS/STOC has been a culture concerned with some sense of "prestige" -- the number of accepted papers has to be kept low, because we want people outside of theory to take FOCS/STOC as an imprimatur for the top theory work.  Becau ..read more

 I have a loose association with the Broad Institute, an institute created so that "complementary expertise of the genomic scientists and the chemical biologists across MIT and Harvard be brought together in one place to drive the transformation of medicine with molecular knowledge."  (See https://www.broadinstitute.org/history ) They recently passed an amazing milestone, having performed over 1 million Covid tests.  They weren't set up to be a Covid testing lab, but the converted their institute space to respond to the Covid crisis.  (See https://covid19-tes ..read more