# What signals sent by author lists

I recently had a conversation about ordering of author lists for papers. Of course, each field has its own conventions but as people start publishing in multiple communities’ venues things can get a bit murky. There are pros and cons and different people have different values, etc.

This is all standard and has been hashed to death.

But what happens when you merge two papers with different author lists? Alphabetical makes things very easy, but if you go a different route, then the primary authors have to slug it out to see who gets first author credit. To split the difference, you could put a footnote saying that authors are in alphabetical order. In the conversation, it came up that putting the footnote implies that there was some tension between the two author groups and so this was the compromise solution after a debate. That was new to me: is this the correct inference to make in most cases?

NB: don’t take this as a sign that I’ve brought the blog back to life for real. I’ve made too many unfulfilled promises on that front…

There was a talk at UChicago in the fall (I’m still on the seminar mailing lists) given by Jan Hązła based, I think, on a paper titled “The Probability of Intransitivity in Dice and Close Elections” with Elchanan Mossel, Nathan Ross, and Guangqu Zheng. The abstract was quite interesting and led to a discussion with my colleagues Emina Soljanin and Roy Yates in which I realized I didn’t quite get the result so I promised to come back after reading it more carefully. Fast forward several months and now I am in Berkeley (on a pre-tenure sabbatical) and they are back east so I figured I could at least blog about it.

The problem is about intransitive dice, which was a new term for me. Consider an $n$-sided die with numbers $\mathbf{a} = (a_1, a_2, a_n)$ and call $\sum_{i=1}^{n} a_i$ the face sum. The die is fair, so the expected face value is $\frac{1}{n} \sum_{i=1}^{n} a_i$. We can define an ordering on dice by saying $\mathbf{a} \succ \mathbf{b}$ if a uniformly chosen face of $\mathbf{a}$ is larger than a uniformly chosen face of $\mathbf{b}$. That is, if you roll both dice then on average $\mathbf{a}$ would beat $\mathbf{a}$.

A collection of dice is intransitive if the relation $\succ$ based on dice beating each other cannot be extended to a linear order. The connection to elections is in ranked voting — an election in which voters rank candidates may exhibit a Condorcet paradox in which people’s pairwise preferences form a cycle: A beats B, B beats C, but C beats A in pairwise contests. (As an aside, in election data we looked at in my paper on instant runoff voting we actually rarely (maybe never?) saw a Condorcet cycle).

Suppose we generate a die randomly with face values drawn from the uniform distribution on $[-1,1]$ and condition on the face sum being equal to $0$. Then as the number of faces $n \to \infty$, three such independently generated dice will become intransitive with high probability (see the Polymath project).

However, it turns out that this is very special to the uniform distribution. What this paper shows (among other things) is that if you generate the faces from any other distribution (but still condition the face sum to be 0), the dice are in fact transitive with high probability. This to me is interesting because it shows the uniform distribution as rather precariously balanced — any shift and a total linear order pops out. But this also makes some sense: the case where the dice become intransitive happens when individuals essentially are choosing a random permutation of the candidates as their preferences. In face, the authors show that if you generate voters/votes this way and then condition on the election being “close” you get a higher chance of turning up a Condorcet paradox.

The details of the proofs are a bit hairy, but I often find ranking problems neat and interesting. Maybe one day I will work on them again…

# PLoS One and its absurdly short review times

I was asked to review a manuscript for PLos One recently and declined because they asked for a review in 10 days. This might be standard for biology papers or something, but seems absurd for a paper where the reviewer is asked to sign off on technical correctness for something which may entail a fair bit of math. This sort of one-size-fits-all approach to academic practice drives me nuts. It’s the same kind of thing that leads to workshops on grant writing led by someone who has had a lot of success writing grants to one program at NIH/NSF/wherever and then dispenses advice specific to that area with almost zero recognition that different programs/agencies have different priorities. Wow, context matters! Who knew?

Now, a reasonable claim is that 10 days at 8 hours a day is 80 hours and that is a totally reasonable amount of time to check all the math in a paper, assuming I had nothing else to do with my time. A friend told me their advisor had a policy to decline a review if they couldn’t do it in the next week. This strikes me as an admirable approach to things that probably worked well in the 80s.

However, given that 50% of papers are accepted to PLoS on a pay-to-publish model, what is the prior belief that spending even 30 minutes of my time reading the paper is worthwhile? Far better to spend 10 minutes complaining about it on a nearly defunct blog, no?

# We’re Hiring! Rutgers ECE, that is.

My department has several openings across a variety of areas. The direct link to apply is here.

The Department of Electrical and Computer Engineering in the School of Engineering at Rutgers, The State University of New Jersey anticipates multiple openings in Fall 2019 for full-time, academic year, tenure-track faculty appointments at a rank and salary commensurate with the applicant’s background and experience.

Diversity and inclusion are a key foundational element of the University’s Strategic Plan and the School of Engineering is fully committed to building a faculty of individuals from diverse backgrounds.  Diversity may include, but is not limited to, gender, ethnicity, race, culture, national origin, or other underrepresented personal or professional characteristics. Applicants who can contribute to this diversity through their teaching, research, or service are particularly encouraged to apply.

Hiring areas for this search are: (icomputer vision, including recognition, machine learning, deep learning,  multi-view 3D reconstruction, computational photography, and medical image analysis; (iiautonomous systems and robotics, including but not limited to networked control systems,  learning and control in autonomous systems such as vehicles or drones as well as in assistive technologies, and human-robot interaction;  (iii) high performance computing (HPC), including but not limited to system software, numerical libraries, middleware, performance engineering, modeling and simulation, HPC applications such as computational medicine, and the intersection of HPC and data-intensive computing.

Exceptional candidates in the university strategic areas are also welcome to apply (see http://rci.rutgers.edu/~presiden/strategicplan/UniversityStrategicPlan.pdf) and will be considered for the newly endowed Rodkin-Weintraub Chair in the School of Engineering.

Excellent facilities are available for collaborative research opportunities with local industry through the School’s nationally recognized centers such as the Wireless Information Network Laboratory (WINLAB), the Microfabrication Laboratory, and the Center for Advanced Infrastructure and Transportation (CAIT) as well as university centers. There also exist several opportunities to collaborate with clinicians at Rutgers University. Rutgers Biomedical and Health Sciences (RBHS) is home to the Center for Advanced Biotechnology and Medicine as well as Rutgers School of Public Health. The Robert Wood Johnson University Hospital, the flagship hospital of the Robert Wood Johnson Health System, is located a few miles from the ECE Department.

A Ph.D. in Electrical and Computer Engineering, Computer Science or a related field is required. All new members of the School of Engineering faculty are expected to develop a research program and obtain external funding to support it.  Candidates will work well in a highly collaborative and interdisciplinary environment and be willing to engage in existing and future large collaborative research endeavors.  A commitment to excellence in teaching at both undergraduate and graduate levels is expected. Demonstrated ability in written and oral use of the English language is required. Candidates must demonstrate a commitment to diversity.

Qualified candidates should submit a CV, statements on teaching and research, and contact information for three references to the URL: http://jobs.rutgers.edu/postings/76931

Questions may be directed to: Narayan Mandayam, Distinguished Professor and Chair, Department of Electrical and Computer Engineering, Rutgers University, narayan@winlab.rutgers.edu.. For full consideration for Fall 2019 openings, applications must be received by January 15, 2019.

EEO/AA Policy: Rutgers is an Equal Opportunity / Affirmative Action Employer. Rutgers is also an ADVANCE institution, one of a limited number of universities in receipt of NSF funds in support of our commitment to increase diversity and the participation and advancement of women in the STEM disciplines.

# DIMACS Director Position!

One of the real delights at Rutgers is being able to work with the DIMACS Center. In particular, I’ve worked closely with Prof. Rebecca Wright, who is the current director of DIMACS. Unfortunately for us, she is moving to Barnard to start a new CS program. So… now we need a new director! Here is the ad…

The DIMACS Center based at Rutgers University in New Brunswick, NJ is seeking a Director.  This is a five-year renewable appointment as Director as well as a tenured full professor position in an appropriate Rutgers department.

Founded in 1989 in the first class of NSF Science and Technology Centers, DIMACS catalyzes and conducts research and education in mathematical, computational, and statistical methods, algorithms, modeling, and data analysis, as well as their applications in the physical, biological, and social sciences and various engineering disciplines.  DIMACS’s multi-year special focus programs address research themes that require topical expertise in these areas, have potential for societal impact, and are poised for advance. Its education programs include materials development for high school and college classrooms, an extensive summer undergraduate research program, and teacher education programs. DIMACS also leads the CCICADA Center for Advanced Data Analysis, which was founded as a Department of Homeland Security University Center of Excellence.  Based at Rutgers, DIMACS’s partners are Columbia, Georgia Tech, Princeton, RPI, Stevens, AT&T, Avaya, IBM, Microsoft, NEC, Nokia, and Perspecta. One of DIMACS’s strengths is its ability to modify its areas of emphasis and programs as times change and interests of its internal and external partners change, and the new Director will be expected to provide leadership and vision in determining such new directions.

The Director will:

• guide the scientific and educational vision of the center;
• lead the center’s efforts to attract funding from public and private sources to support the growth and long-term sustainability of the center;
• identify and pursue opportunities for building research and educational collaborations across the Rutgers faculty, with our partners, and with others

Candidates must have a strong research record suitable for a tenured appointment and leadership position at Rutgers in a field such as mathematics, computer science, operations research, statistics, data science, or applications of these disciplines to other areas of science and/or engineering. The candidate should display breadth of scientific and educational interests and have demonstrated leadership abilities. A successful record of grant funding is strongly preferred.

Interested candidates should submit a current Curriculum Vitae and a cover letter describing their interest in the position to Lazaros Gallos and he will inform them of the formal application procedure once it is set up. The initial application review will begin on January 15, 2019 and will continue until the position is filled. No letters of reference are required during the initial stage of the search process.

# Some thoughts on teaching signals and systems

I’m teaching Linear Systems and Signals[*] (ECE 345) this semester at Rutgers. The course overall has 260+ students, split between two sections: I am teaching one section. This is my second time teaching it: last year I co-taught with Vishal Patel (who has decamped to Hopkins), and this semester I am co-teaching with Sophocles Orfanidis. I inherited a bit of a weird course: this is a 3-unit junior-level class with an associated 1-unit lab (ECE 347). Previous editions of the course had no recitations, which boggled my mind, since the recitation was where I really learned the material when I took the course (6.003 at MIT, with Greg Wornell as my recitation instructor). How are you supposed to understand how to do all these transforms without seeing some examples?

So this year we have turned ECE 347 into a recitation and moved the coding/simulation part of the course into the homework assignments. Due to the vagaries of university bureaucracy, however, we still have to assign a separate grade for the recitation (née lab). Moreover, there are some students who took the class without the lab and now just need to take 347! It’s a real mess. Hopefully it’s just one year of transition but this is also the year ABET [**] is showing up so we’ll see how things go.

After surveying a wide variety of textbook options for the course, we decided to go with the brand-new and free book by Ulaby and Yagle, Signals and Systems: Theory and Applications [***]. I really have to commend them on doing a fantastic job and making the book free, which is significantly better than \$247 for the same book I used literally 20 years ago when I took this course. Actually, we mainly used another book, whose title/author eludes me now, but it had a green slipcover and was more analog control-focused (perhaps since Munther Dahleh was teaching).

One major difference I noticed between textbooks was the order of topics. Assuming you want to do convolution, Laplace (L), Z, Fourier Series (FS), and Fourier Transforms (FT), you can do a sort of back and forth between continuous time (CT) and discrete time (DT):

CT convolution, DT convolution, CTFS, DTFS, CTFT, DTFT, Laplace, Z
CT convolution, DT convolution, Laplace, Z, CTFS, DTFS, CTFT, DTFT

or do all one and then the other

CT convolution, Laplace, CTFS, CTFT, DT convolution, Z, DTFS, DTFT
DT convolution, Z, DTFS, DTFT, CT convolution, Laplace, CTFS, CTFT

I like the alternating version because it emphasizes the parallels between CT and DT, so if you cover sampling at the end you can kind of tie things together. This tends to give students a bit of whiplash, so we are going for:

CT convolution, DT convolution, Laplace, Z, CTFS, CTFT, DTFS, DTFT

It’s all a bit of an experiment, but the thing I find with all textbooks is that they are never as modular as one might like. That’s good for a book but maybe not as good for a collection of curricular units, which in the end is what a S & S [****] class is. CNX is one type of alternative, or maybe something like the interactive book that my colleague Roy Yates dreams of.

I find myself questioning my own choices of ordering and how to present things in the midst of teaching — it’s tempting to experiment mid-stream but I have to tamp down the urges so that I don’t lose the class entirely.

[*] You can tell by the word ordering that it was a control theorist who must have named the course.

[**] Accreditation seems increasingly like a scam these days.

[***] You can tell by the word ordering where the sympathies of the authors lie.

[****] Hedging my bets here.

# CFP: PPML Workshop at NIPS 2018

### Montreal, December 8, 2018

Description

This one day workshop focuses on privacy preserving techniques for training, inference, and disclosure in large scale data analysis, both in the distributed and centralized settings. We have observed increasing interest of the ML community in leveraging cryptographic techniques such as Multi-Party Computation (MPC) and Homomorphic Encryption (HE) for privacy preserving training and inference, as well as Differential Privacy (DP) for disclosure. Simultaneously, the systems security and cryptography community has proposed various secure frameworks for ML. We encourage both theory and application-oriented submissions exploring a range of approaches, including:

• secure multi-party computation techniques for ML
• homomorphic encryption techniques for ML
• hardware-based approaches to privacy preserving ML
• centralized and decentralized protocols for learning on encrypted data
• differential privacy: theory, applications, and implementations
• statistical notions of privacy including relaxations of differential privacy
• empirical and theoretical comparisons between different notions of privacy
• trade-offs between privacy and utility

We think it will be very valuable to have a forum to unify different perspectives and start a discussion about the relative merits of each approach. The workshop will also serve as a venue for networking people from different communities interested in this problem, and hopefully foster fruitful long-term collaboration.

Submission Instructions

Submissions in the form of extended abstracts must be at most 4 pages long (not including references) and adhere to the NIPS format. We do accept submissions of work recently published or currently under review. Submissions should be anonymized. The workshop will not have formal proceedings, but authors of accepted abstracts can choose to have a link to arxiv or a pdf published on the workshop webpage.

Program Committee

• Borja de Balle Pigem (Amazon)
• Emiliano de Cristofaro (University College London)
• David Evans (University of Virginia)
• Irene Giacomelli (Wisconsin University)
• Nadin Kokciyan (King’s College London)
• Kim Laine (Microsoft Research)
• Payman Mohassel (Visa Research)
• Catuscia Palamidessi (Ecole Polytechnique & INRIA)
• Mijung Park (Max Planck Institute for Intelligent Systems)
• Benjamin Rubinstein (University of Melbourne)
• Anand Sarwate (Rutgers University)
• Philipp Schoppmann (HU Berlin)
• Nigel Smart (KU Leuven)
• Carmela Troncoso (EPFL)
• Pinar Yolum (Utrecht University)
• Samee Zahur (University of Virginia)

Organizers

• Adria Gascon (Alan Turing Institute & Edinburgh)
• Niki Kilbertus (MPI for Intelligent Systems & Cambridge)
• Olya Ohrimenko (Microsoft Research)
• Mariana Raykova (Yale)
• Adrian Weller (Alan Turing Institute & Cambridge)

# Signal boost: vote in the ITSOC BoG Election!

The IEEE Information Theory Society election of Members to the Board of Governors is now open.

We hope you will take the time to exercise your vote and help choose the future direction of the society. Let’s increase last year’s voting percentage of 23.1% by voting today! Your vote counts!

VOTE NOW at https://eballot4.votenet.com/IEEE

Voting must be completed no later than 6 September 2018. Any returns received after this date will not be counted. The online voting site will close at 4:00 pm Eastern Time.