DIMACS Workshop on Network Coding: the Next 15 Years

DIMACS Workshop on Network Coding: the Next 15 Years
December 15 – 17, 2015
DIMACS Center, CoRE Building, Rutgers University

Organizers:

  • Michael Langberg, SUNY Buffalo
  • Emina Soljanin, Bell Labs, emina at research.bell-labs.com
  • Alex Sprintson, Texas A&M

Presented under the auspices of the Special Focus on Cybersecurity, the Special Focus on Information Sharing and Dynamic Data Analysis and the Special Focus on Energy and Algorithms.

Since its introduction at the turn of the millennium, network coding has evolved from a simple idea to a mature interdisciplinary filed, with a solid body of knowledge generated by a diverse group of researchers. Over the years, network coding techniques have had significant impacts and benefits in throughput, reliability, security, and energy efficiency. We believe that in the next fifteen years the field will expand even further into new multidisciplinary area and provide a fertile ground for the next generation of researchers, leading to new breakthroughs, discoveries, and solutions to long standing open problems.

The goal of this workshop is to discuss the long-term horizons of the field and identify key areas and research problems that will be in the focus of the research community. The workshop will include speakers from a broad spectrum of backgrounds, from theoreticians to practitioners, and from the founders of the field to younger faculty and students.

This is list of Confirmed Participants:

  • Alexander Barg, UMD
  • Eimear Byrne, University College Dublin
  • Viveck Cadambe, Penn State
  • Chandra Chekuri, UIUC
  • Hoang Dau, UIUC
  • Alex Dimakis, UT Austin
  • Michelle Effros, Caltech
  • Salim El Rouayheb, Illinois Institute of Technology
  • Sid Jaggi, CUHK
  • Shirin Jallali, Bell Labs
  • Sudeep Kamath, Princeton
  • Young-Han Kim, UCSD
  • Joerg Kliewer, New Jersey Institute of Technology
  • Oliver Kosut, Arizona State University
  • Gerhard Kramer, TU Munich
  • Michael Langberg, SUNY Buffalo
  • Mohammad Ali Maddah-Ali, Bell Labs
  • Muriel Medard, MIT
  • Bobak Nazer, Boston University
  • Aditya Ramamoorthy, Iowa State University
  • Parastoo Sadeghi, Australian National University
  • Emina Solanin, Bell Labs
  • Alex Sprintson, Texas A&M
  • Raymond Yeung, CUHK

Call for Participation:

Attendance at the workshop is open to all interested participants (subject to space limitations). Please register if you would like to attend this workshop.

Family leave for graduate students: how does it work at your school?

I am trying to understand how family leave works for graduate students at different schools. More specifically, I am interested in how the finances for family leave work. Graduate students at Rutgers (as at many schools) are covered by a union contract. The contract specifies that in case of a pregnancy, the mother can take 6 weeks of paid leave recovery time plus an additional 8 weeks of paid leave family time. Non-carrying parents can take 8 weeks of paid leave for family time. While not generous by European standards, it’s better than what I would expect (ah, low expectations) here in the US.

This raises the question of how the university pays for the leave time. Students are either teaching or research assistants. 14 weeks off from teaching might include most of a semester, so the department needs a substitute. Trying to give the student an “easy TA” and still expecting them to come and teach when they are entitled to the leave is shady (although I have heard this idea floated). If they are paid through a grant, how should the leave time be charged?

I recently contacted authorities at Rutgers about this, and their response was not encouraging. Rutgers foists all charges off onto the department or grant/PI. If you are a TA and have a baby, the department is on the hook, financially, for finding a replacement. If you are a research assistant, they just charge the paid leave to the grant, as per the fringe rules in OMB Circular A-21.

I wrote a letter back about how disappointing this all is. The current system creates strong incentives for departments and PIs to deny appointments to students who have or may develop family obligations. This lack of support from the University could result in systematic discrimination against student parents. Whether examples of such discrimination exist is not clear, but I wouldn’t be surprised. Allocating the financial burden of leave to departments creates great inequities based on department size and budget, and not all departments can “close ranks” so easily.

For PIs covering students on grants with “deliverables,” the system encourages not supporting students on such grants. The rules in OMB Circular A-21 say that costs should be “distributed to all institutional activities in proportion to the relative amount of time or effort actually devoted by the employees.” It also implies that leave time should be charged via fringe benefits and not salary. It’s not entirely clear to be how a particular grant should be charged if a student participant goes on family leave, but the Rutgers policy seems to be to stick it to the PI.

The current situation leaves students in a predicament: when should they tell their advisor or department that they are pregnant? Many students are afraid of retribution or discrimination: I have heard from students that their friends say advisors “don’t like it when their students have kids.” The university’s policy on this issues only serves to legitimize these fears by creating uncertainty for them about whether they will be reappointed.

My question to the readers of this blog is this: how does your university manage paying for family leave for grad students?

LabTV, research stories, and video outreach

My lab was visited by Charlie Chalkin a few weeks ago. He was here to interview me and various students on our experiences in research for LabTV. LabTV was founded by Jay Walker and the NIH director Dr. Francis Collins with the aim of profiling NIH-funded researchers (as I now am). It was a great opportunity and a really short informal process, and I guess I can get some more hits from YouTube on the LabTV channel.

This experience got me thinking about how hard it is to connect with students at times. In particular, I think that many students don’t really see the process of how we got to where we are as their professors. Unless they have an academic in the family and also paid attention to their life story, they seem a bit mystified by it all. Obviously pop culture has a lot to do with this — movie and TV depictions of the professoriat are pretty far from reality. I have heard, however, from Ram Rajagopal that San Andreas has pretty much the most amazing interactions between professors and grad students. Heroism — that’s what we want.

But this experience got me thinking that departments might benefit from having short 2 minute profiles of their faculty members, but not from the technical achievements view. Instead, let them talk about what got them interested in the problems they are interested in, how they ended up in this position, and why they like the job. The answers may be surprising, but I think students might see a different side than they get in the lecture hall.

Call for Papers: T-SIPN Special Issue on Distributed Information Processing in Social Networks

IEEE Signal Processing Society
IEEE Transactions on Signal and Information Processing over Networks
Special Issue on Distributed Information Processing in Social Networks

Over the past few decades, online social networks such as Facebook and Twitter have significantly changed the way people communicate and share information with each other. The opinion and behavior of each individual are heavily influenced through interacting with others. These local interactions lead to many interesting collective phenomena such as herding, consensus, and rumor spreading. At the same time, there is always the danger of mob mentality of following crowds, celebrities, or gurus who might provide misleading or even malicious information. Many efforts have been devoted to investigating the collective behavior in the context of various network topologies and the robustness of social networks in the presence of malicious threats. On the other hand, activities in social networks (clicks, searches, transactions, posts, and tweets) generate a massive amount of decentralized data, which is not only big in size but also complex in terms of its structure. Processing these data requires significant advances in accurate mathematical modeling and computationally efficient algorithm design. Many modern technological systems such as wireless sensor and robot networks are virtually the same as social networks in the sense that the nodes in both networks carry disparate information and communicate with constraints. Thus, investigating social networks will bring insightful principles on the system and algorithmic designs of many engineering networks. An example of such is the implementation of consensus algorithms for coordination and control in robot networks. Additionally, more and more research projects nowadays are data-driven. Social networks are natural sources of massive and diverse big data, which present unique opportunities and challenges to further develop theoretical data processing toolsets and investigate novel applications. This special issue aims to focus on addressing distributed information (signal, data, etc.) processing problems in social networks and also invites submissions from all other related disciplines to present comprehensive and diverse perspectives. Topics of interest include, but are not limited to:

  • Dynamic social networks: time varying network topology, edge weights, etc.
  • Social learning, distributed decision-making, estimation, and filtering
  • Consensus and coordination in multi-agent networks
  • Modeling and inference for information diffusion and rumor spreading
  • Multi-layered social networks where social interactions take place at different scales or modalities
  • Resource allocation, optimization, and control in multi-agent networks
  • Modeling and strategic considerations for malicious behavior in networks
  • Social media computing and networking
  • Data mining, machine learning, and statistical inference frameworks and algorithms for handling big data from social networks
  • Data-driven applications: attribution models for marketing and advertising, trend prediction, recommendation systems, crowdsourcing, etc.
  • Other topics associated with social networks: graphical modeling, trust, privacy, engineering applications, etc.

Important Dates:

Manuscript submission due: September 15, 2016
First review completed: November 1, 2016
Revised manuscript due: December 15, 2016
Second review completed: February 1, 2017
Final manuscript due: March 15, 2017
Publication: June 1, 2017

Guest Editors:

Zhenliang Zhang, Qualcomm Corporate R&D (zhenlian@qti.qualcomm.com)
Wee Peng Tay, Nanyang Technological University (wptay@ntu.edu.sg)
Moez Draief, Imperial College London (m.draief@imperial.ac.uk)
Xiaodong Wang, Columbia University (xw2008@columbia.edu)
Edwin K. P. Chong, Colorado State University (edwin.chong@colostate.edu)
Alfred O. Hero III, University of Michigan (hero@eecs.umich.edu)

Readings

Mirrors (Naguib Mahfouz): This is a collection of character sketches that were serialized in a magazine that Mahfouz wrote for. Each chapter is a different person — they are related, living through the same era in Cairo, but there is not a story here. Collectively they evoke a sense of how much Egypt changed between independence and 1970, and I found myself having to look up a lot of things, like the history of the Wafd party. If you already like Mahfouz or you’re interested in literary takes on mid-20th century Egypt, then it’s worth reading.

Rebetiko (David Prudhomme): Set in 1936 Athens, this graphic novel is about a day in the life of a group of rebetiko musicians. The dictatorship is on the rise, and fascism has no room for degenerate music like rebetiko. There is a slight arc to the story and links to specific historical figures, but it’s more about the time and the place and the music. My quondam co-blogger and co-author Alex is in a rebetiko band down in Austin.

Odysseus Abroad (Amit Chaudhuri): A recasting of Ulysses, a bit, starring Ananda, a poetry student from India going to university in London, and his uncle, who lives in a bedsit in Belsize Village after years of working in the accounting end of a maritime shipping business. They wander, eat, and talk. Ananda confronts his own sense of out-of-place-ness. He cares deeply for his uncle, who is still a bit frustrating. Both are alienated and the similarities and differences, carefully observed, are what make this novel worth reading.

Oreo (Fran Ross): I was on a Greek-inpsired modern fiction kick, apparently. Oreo is a satire starting a half-black half-Jewish kid named Oreo who goes on a Thesus-like journey to find her father. The book is full of allusions to the myth (there’s a helpful gloss in the back) and is sprinkled with many Yiddishisms. It’s received a bit of a resurgence in interest because it confronts these issues of hybridity and identity, although the aesthetic is rooted in a 70s broad farcical style that reminded me of Ishmael Reed. I want more people to read it so that I can talk to them about it — so much to discuss!

The Philosopher Kings (Jo Walton): a sequel to Walton’s The Just City, set several years later. If you liked the first book you might like this one too, but it’s definitely more of a “more adventures in the land of.” While it addresses some more weighty topics such as revenge and power/divinity and other knotty philosophical issues, I didn’t find it as “surprising” as the first book. There are some lovely scenes in there and Walton’s deftness at switching between different first-person narrators is really a delight. The structure of the story feels almost… architectural (but not in a David Mitchell way). Recommended if you liked the first.

Mathematical Tools of Information-Theoretic Security Workshop: Days 2-3

I took sketchier notes as the workshop progressed, partly due to the ICASSP deadline, but also because jet lag started to hit me. The second day was a half day, which started with Zhenjie Zhang giving a tutorial on differential privacy from a databases/data mining perspective and my talk on more machine learning aspects. In between us was a talk by Ben Smyth on building automatic verification for security protocols. Basically you write the protocol as a program and then the ProVerif verifier will go and try to break your protocol. As an example, it can automatically find/generate a man-in-the-middle attack if one exists. I thought it was pretty neat, especially after having recently talked to someone about automatic proof systems. It’s based on something called the applied pi calculus, which I did not understand at all, but hey, I learned something new, which was great. The last two talks of the day were by Lalitha Sankar and Mari Kobayashi. Lalitha talked about mutual information based measures of privacy leakage in an interactive communication setting that is the information-theoretic analogue of communication complexity models in CS. Mari talked about the broadcast channel with state feedback. This is trying to find secure analogues of these opportunistic multicast settings where you need to also generate a secret key.

The last day was on quantum! I learned a lot and took few notes, unfortunately. Andreas Winter gave a tutorial on quantum (the slides for most talks are online and his are as well) and Ciara Morgan discussed the challenges in proving a strong converse for the the capacity of quantum channels. Damian Markham talked about secret sharing in quantum systems. Masahito Hayashi gave a very densely-packed talk surveying a large number of results based on secure randomness extraction and hash functions using Rényi information measures. I think privacy amplification is really interesting but I think I need a tutorial on it before I can really get the research results. The last non-overview talk I have notes on was by David Elkouss (apologies to the remaining speakers): this was a really interesting presentation on how to decide which of two channels is better from a quantum communication sense. The slides are a little engimatic, but the papers are online.

Shlomo Shamai made it to the last day of the workshop (the intersection with High Holidays was unfortunate) — he talked about the layered secrecy view of the broadcast channel: rather than thinking only of the secret message as carrying information, one can think of certain layers (c.f. superposition coding) as being secured based on the channel to the non-legitimate receiver. For example, in a degraded broadcast channel, the strong receiver’s message can sometimes be thought of as secret from the weak receiver. This leads to a raft of models and setups based on who wants to keep what secret from whom, shedding some light on standard superposition, rate splitting, binning, and embedding constructions. The talk was largely based on a paper in the current issues of the Proceedings of the IEEE.

All in all, this was a really great workshop, and the organizers were very generous in the organization.