Some pictures from ICALP

On 10-13 July we hosted ICALP in Warsaw. It was fun! See a few random pictures below.

Invitation to Award Ceremony of the ESA Test-of-Time Award 2016

On behalf of the Steering Committee of ESA and the Program Committee of ESA 2017, I would like to invite you to attend the ESA conference and the ESA Test-of-Time 2016 Award Ceremony. The ceremony will take place on the afternoon of 5th of September during ESA 2017 in Vienna and it will include the presentation by the awardees of the recognized paper.
Announcement of the ESA Test-of-Time Award 2016

European Symposium on Algorithms (ESA)

The ESA Test-of-Time Award (ESA ToTA) recognizes excellent papers in algorithms research that were published in the ESA proceedings 19-21 years ago and which are still influential and stimulating for the field today. In this second year in which the award is given, papers from ESA'95 to ESA'97 were considered.

The committee nominates the following paper for the ESA ToTA 2016. The paper stands out as a classic in the algorithms field and by its excellent citation
record still relevant today.

From ESA 95-97:

Boris V. Cherkassky, Andrew V. Goldberg
Negative-cycle detection algorithms
Proceedings ESA'96, also in: Mathematical Programming 85:2 (1999) 277-311

The paper by Cherkassky and Goldberg deals with the problem of finding a
negative-length cycle in a network or proving that there is none. Algorithms
for this are a combination of a shortest-path algorithm and a negative-cycle
detection strategy. The authors analyse known algorithms and some new ones
and determine the best combinations. Novel instance generators are used in
this study. The paper is a model experimental paper in algorithms.

ESA ToTA 2016 Award Committee:
Kurt Mehlhorn (Saarbrucken)
Mike Paterson (Warwick)
Jan van Leeuwen (Utrecht)

We in Warsaw, recently stumbled upon which ranks CS departments world-wide according to publication counts in mayor conferences. A cool aspect of this ranking is that one can set the year range for the counted publications. We rather only very recently, i.e., during the last 5-10, started doing well, so this parameter allows to track this change. As we are working mostly in theory let us restrict the ranking to theory only, i.e, Algorithms & complexity; Cryptography; Logic and verification. If you set the range to start in 2000 - to contain the last 15 years, then we are placed on position 10 world-wide. Quite cool already. Setting the starting point to 2005 bumps us to the 6th position. The most interesting part happens when you set the starting point to 2010, i.e., last 5 years, as you get the following table:

Rank Institution Average Count Faculty
1 ? University of Warsaw ? 8.4 24
2 ? Massachusetts Institute of Technology ? 8.2 20
3 ? University of Texas - Austin ? 7.3 8
4 ? Stanford University ? 7.1 12
5 ? Cornell University ? 6.9 14
6 ? Northeastern University ? 6.1 10
7 ? New York University ? 5.9 10
7 ? Carnegie Mellon University ? 5.9 14
9 ? University of Illinois at Urbana-Champaign ? 5.7 14
9 ? University of California - San Diego ? 5.7 10

Quite surprising... as all the ranking are usually. However, there might be some reasons for this. First of all, we cover all three areas that are counted in theory. For example, logic is visibly stronger in Warsaw than in other places included in top 10. Second, we indeed started doing quite well, e.g., we got 6 ERC grants so far. I do not really believe in rankings, but at least it is some indication that Warsaw is not such a bad place to be - see the cat meme.

Bratislava Declaration of Young Researchers

The Bratislava Declaration of Young Researchers is something I was involved in recently. Its preparation was inspired by Slovak Presidency of the EU and it was presented on today's informal Council of Ministers responsible for competitiveness (Research). I hope this will have some follow up, as current trend in funding research in EU is in my opinion (and not only my as this declaration shows) going in the wrong direction.

Why should you not use the automatic assignment by EasyChair

Some days ago I have assigned papers to my PC in ESA Track A.  I tried to use automatic assignment in EasyChair and this was a rather disappointing experience. Just by running it I understood why some assignments I got as a PC member in previous conferences were so bad - they bad without any good reason. Let me explain what happened this time.

I clicked the automatic assignment button and I got a very unfair assignment. EasyChair wanted to punish 3 of my PC members by giving them 11, drugs 12 and 15 papers from their No list. In EasyChair one can bid Yes, Maybe and No on papers. I found this rather absurd as 15 paper was more than half of the total review load. The question is: Does this poor guy need to get so many Nos in optimal assignment?. The short intuitive answer is: most probably no, as there is a lot of freedom in each maximal size matching (by Gallai-Edmonds decomposition).

Hence, in 2 hours I coded my own assignment procedure. I assumed that in the assignment:

  • first, I want to maximize the number of papers assigned to Maybes and Yeses,
  • second, under the first assumption, I want to maximize the number of papers assigned Yeses.

This can be formalized as a maximum-cost maximum-flow problem:

  • connect the source vertex with each PC member with a zero cost edge of capacity equal to review load,
  • connect each PC member with each paper with edges of capacity 1 and cost 0 for No, 10000 for Maybe and 10001 for Yes.
  • connect each paper with the sink with edge of cost 0 and capacity 3.

I executed it and I got assignment with the same number of Nos - 80 and Maybes -100 as EasyChair. And of course, it was not much more fair than the one given by EasyChair. If you always wondered how EasyChair computes the assignment this is the way. Essentially, there is no reason for this to be fair in any sense. One is optimizing global cost, so it should be clear that locally it can be bad, e.g., some PC members will get bad assignments without any good reason. It can be even worse as one can be getting many Nos due to the execution order of procedure, as there is always a lot of freedom in the optimal matching. It can happen that even if you bid on almost everything you will get all the Nos if all people bid No on the same papers.

There is a solution here - I added third assumption to my procedure:

  • zero, no PC member can get more than X Nos, where X is a parameter.

This can be easily incorporated into the max-flow problem. One just needs to split a vertex representing each PC member into three vertices: PC_0, PC_No and PC_NotNo. The maximum-cost maximum-flow problem becomes:

  • connect the source vertex with all PC_0 vertices with a zero cost edge of capacity equal to review load,
  • connect PC_0 to PC_No with zero cost edge of capacity X,
  • connect PC_0 to PC_NotNo with zero cost edge of capacity equal to review load,
  • connect PC_No with edges of cost 0 to all papers with No,
  • connect PC_NotNo with each paper with edges of capacity 1 and cost 10001 for Maybe and 10001 for Yes.
  • connect each paper with sink with edge of cost 0 and capacity 3.

I executed it - the table below shows the number of Maybes in the assignment in dependence on the parameter X.

 X 4 5 6 7 8 9 10 11 12
 #Maybes 120 116 112 109 106 104 102 100 100

One striking thing is that there exists an optimal assignment where the poor guy that was getting 15 Nos is getting only 11 Nos. Essentially, there was no reason for giving 15 Nos to this PC member! This probably already happened to many of you ;). I the final assignment I went for X=4 as this actually meant that almost everyone gets 3 Nos and only 2 PC members get 4 Nos. This much fairer assignment costed only 20 Maybes in the quality, what I found a fair cost of being fairer.

Highlights of Algorithm Conference - Registration is Open

The HALG conference has opened the registration and the program is available as well - see the links below. The program looks very interesting so I hope many of you will be coming.

Highlights of Algorithms - HALG 2016
June 6-8, link 2016, ask Paris, France

The Highlights of Algorithms conference is designed to be a forum for presenting the highlights of recent developments in algorithms and for discussing potential further advances in this area. The conference will provide a broad picture of the latest research in algorithms through a series of survey and invited talks, as well as possibility for all researchers and students to present their recent results through a series of short talks and poster presentations. Attending the Highlights of Algorithms conference will also be an opportunity for networking and meeting leading researchers in algorithms.

The registration for the Highlights of Algorithms conference is open.
To register, please visit the link at (early registration is due April 30)

A preliminary program is available online at
The program is packed with 28 invited talks and with even a larger number of short contributions.

Those interested in attending the conference are advised to book accommodation as early as possible (given the high hotel prices in Paris).


ESA 2016 - call for papers

The ESA call for papers is out: The submission deadline is April 21, 23:59 AoE, see 2016, whereas the notifications will be send out no later than on June 9, 2016. I hope there will be many submissions to keep us in the PC busy during this time. Anyway, if you plan to submit something, you might read (or maybe read again) the post by Boaz Barak: It contains advice on how to increase your chances that your paper gets accepted to a conference - FOCS in this case. The same applies to most conferences and "putting the worst foot forward" is often the hardest part. However, believe me that doing the opposite usually hurts even more.

Highlights of Algorithm Conference

Guest post by Aleksander M?dry

This summer, drug on June 6-8 in Paris, help we will be having a new algorithmic event: Highlights of Algorithms 2016 (HALG 2016) conference.

This conference will be quite unlike the conferences we are all used to. First of all, patient it will consist mainly of invited talks and tutorials, accompanied by a smaller number of contributed talks and posters. Secondly, there will be no conference proceedings, i.e., presenting work already published at a different venue or journal (or to be submitted there) is absolutely welcome and even encouraged.

The call for submissions  of contributed short talks and posters has been just posted online. The program committee is currently in the process of selecting invited speakers (partially based on the nominations from the community).

We already have a number of very exciting speakers committed and we expect to add a few more names shortly.

The intention here is to make HALG a forum for presenting the highlights of some of the most exciting recent developments in algorithms, discussing potential further advances in this area, as well as networking and initiating new collaborations. In a way, we want HALG to be a place one can go to catch up on what are the hottest results/topics
at the moment, as well as what one missed by not being able to attend all the STOC/FOCS/SODA/ESA/ICALP/PODC/ITCS/… conferences that year.

We believe that this kind of venue is very much needed today and will help to make our community stronger and more cohesive. In particular, we hope that the success of the first edition will make HALG become a recurring event. In fact, HALG draws some inspiration from Highlights of Logic, Games and Automata (HLGA) conference series that became very successful in its respective community already. There are some differences in how both conferences are constructed: HALG will concentrate on invited talks with some contributed talks, whereas HLGA was only having short contributed talks so far; but the overall premise is the same and seems to be working very well. (We also heard some rumors that the next edition of HLGA will move in the direction of HALG and have invited talks too.)

In any case, the HALG 2016 in Paris should be a lot of fun. We hope to see many of you there! So please submit your work to HALG.

Two ERC grants in our group are starting!

We are quite happy that this year two ERC grants will be starting in our group. First of all, Marek Cygan very recently got news that his ERC Starting Grant received funding. His grant is on FPT, approximation and metaheuristics. Secondly, our ERC Proof of Concept grant, that is a follow up to my ERC StG grant is starting. This grant aims to commercialise some ideas from algorithmic market modeling and to support formation of a spin-off.

On a (Somewhat Old) Dynamic Steiner Tree Problem

The dynamic Steiner tree problem has been around for a while already (for 24 years), but it did not get a satisfying answer from efficiency point of view, i.e., one would like to have a fast algorithm that maintains a constant approximate solution and allows to update the set of terminal vertices. On one hand, there have been some studies of the online version of this problem, where we focus on minimizing the number of changes to the tree that are necessary to maintain a good approximation. This line of research was started in the pioneering paper by Imase and Waxman [1] and was later continued in [2, 3, 4]. However, in the online problem one ignores the problem of efficiently finding these changes to the tree. On the other hand, the problem of maintaining a Steiner tree is also one of the important open problems in the network community [5], where it is often referred to as dynamic multi-cast tree problem. The motivating problem is to maintain a cheap communication network between a set of dynamically changing users. While it has been studied for many years, the research resulted only in several heuristic approaches [6, 7, 8, 9], none of which has been formally proved to be efficient.

In our paper "The Power of Dynamic Distance Oracles: Efficient Dynamic Algorithms for the Steiner Tree" we have shown the first sublinear time algorithm which maintains a constant approximate Steiner tree under terminal additions and deletions. We show that we can maintain a (6+epsilon)-approximate Steiner tree of a general graph in  ilde{O}(sqrt{n} log D) time per terminal addition or removal. Here, D denotes the stretch of the metric induced by the graph. For planar graphs we achieve the same running time and the approximation ratio of (2+epsilon). Observe that due to the sparsity of real world networks only algorithms with such sublinear complexity could be of some practical importance and could potentially be used to reduce the communication cost of dynamic multicast trees. Our paper aims to be a theoretical proof of concept that from algorithmic complexity perspective such algorithms are indeed possible. However, turning these theoretical ideas into useful algorithms will require much more research. In particular, a natural question here is whether the approximation factor of our algorithms could be considerably improved while keeping the running time sublinear?

Let me note that this might be hard as the approximation factor of (6 + epsilon) for Steiner tree in general graphs comes from using 3-approximate oracles [10], using 2-approximation of the Steiner tree by the MST in the metric closure, and (1 + epsilon)-approximate online MST algorithm. In other words we hit two challenging bounds: in order to improve our approximation factors, one would need either to improve the approximation ratio of the oracles which are believed to be optimal, or devise a framework not based on computing the MST. The second challenge would require to construct simple and fast (e.g., near-linear time) approximation algorithms for Steiner tree that would beat the MST approximation ratio of 2. Constructing such algorithms is yet another challenging open problem.

If you would like to learn more on how it works, you can still attend our talk during the coming STOC in Portland on Monday at 9:00, or have look into the full version of the paper on arXiv.

[1] M. Imase and B. M. Waxman. Dynamic Steiner tree problem. SIAM J. Discrete Math., 4(3):369–384, 1991.
[2] N. Megow, M. Skutella, J. Verschae, and A. Wiese. The power of recourse for online MST and TSP. In ICALP, pages 689–700. 2012.
[3] A. Gu, A. Gupta, and A. Kumar. The power of deferral: maintaining a constant-competitive Steiner tree online. In STOC, pages 525–534, 2013.
[4] A. Gupta and A. Kumar. Online Steiner tree with deletions. In SODA, pages 455–467. SIAM, 2014.
[5] X. Cheng, Y. Li, D.-Z. Du, and H. Ngo. Steiner trees in industry. In D.-Z. Du and P. Pardalos, editors, Handbook of Combinatorial Optimization, pages 193–216. Springer US, 2005.
[6] F. Bauer and A. Varma. ARIES: A rearrangeable inexpensive edge-based on-line Steiner algorithm. IEEE Journal of Selected Areas in Communications, 15:382–397, 1995.
[7] E. Aharoni and R. Cohen. Restricted dynamic steiner trees for scalable multicast in datagram networks. In INFOCOM ’97. Sixteenth Annual Joint Conference of the IEEE Computer and Communications Societies. Driving the Information Revolution., Proceedings IEEE, volume 2, pages 876–883 vol.2, Apr 1997.
[8] S.-P. Hong, H. Lee, and B. H. Park. An efficient multicast routing algorithm for delaysensitive applications with dynamic membership. In INFOCOM ’98. Seventeenth Annual Joint Conference of the IEEE Computer and Communications Societies. Proceedings. IEEE, volume 3, pages 1433–1440 vol.3, Mar 1998.
[9] S. Raghavan, G. Manimaran, C. Siva, and R. Murthy. A rearrangeable algorithm for the construction of delay-constrained dynamic multicast trees. IEEE/ACM Transactions on Networking, 7:514–529, 1999.
[10] M. Thorup and U. Zwick. Approximate distance oracles. J. ACM, 52(1):1–24, 2005.