Tuesday, January 31, 2017

Call for expressions of interest for permanent positions at the Gran Sasso Science Institute

I have been asked to distribute this call. The Gran Sasso Science Institute is a new, ambitious international centre for advanced studies and PhD school. Give this opportunity a thought. 

Gran Sasso Science Institute - Computer Science Area

Expression of Interest for Faculty Positions


Gran Sasso Science Institute (GSSI) is a new Center for Advanced studies and PhD School established in L’Aquila (Italy).
The Computer Science Area of GSSI (http://cs.gssi.infn.it/phd-program/information/) invites expressions of interest for permanent positions at the level of Full or Associate Professor, and for tenure track (Ricercatore TD – B), and temporary (Ricercatore TD – A) research assistants. Highly qualified candidates with a strong research background in the following fields: cryptography, network security, algorithms, and software engineering are encouraged to apply. The Institute is particularly interested in (potential) leaders of multi-disciplinary research groups in the above fields.
Candidates must have an excellent record of publications, a clear potential to promote and lead research activities, and a specific interest in teaching at the postgraduate level to skilled students recruited internationally.

APPLICATIONS
Applicants should submit their expression of interest by sending
  • a motivation letter in which also the position(s) of interest are specified.
  • a curriculum vitae
  • a list of publications,
  • a brief research statement.
Applications (and questions regarding the application process) must be submitted in electronic form, preferably by April 1st 2017, to eoi-cs@gssi.it.

DISCLAIMER
Please note that this is not a permanent job vacancy advertisement. The expressions of interest received will contribute to the decision of the Gran Sasso Science Institute whether or not to open official calls, their number and the selection procedures.

ADDITIONAL INFORMATION

1- Duties
Teaching post-graduate courses, leading internal seminar and tutoring PhD students. All activities are in English.
2- Salary
The salary will be determined on a personal basis, also taking into account past positions covered abroad. Indicative figures of the gross amount for starting (minimum) salary are:
  • Professor (prima fascia) € 72.431
  • Associate Professor (seconda fascia) € 50.831
  • Assistant Professor, tenure track (ricercatori TD – B) € 34.898
  • Assistant Professor, no tenure track (ricercatori TD – A) € 34.898
Net income may vary depending on income taxes, local taxes, retirement plan, health care deduction and tax exemptions. Professors who have held a tenured position outside Italy for more than three years (at the corresponding level) might be eligible for a partial recognition of past services, depending on specific legal constraints.

Tuesday, January 24, 2017

EATCS Award 2017 to Éva Tardos

The EATCS is proud to announce that the EATCS Award Committee consisting of Fedor Fomin (chair), Christos Papadimitriou and Jean-Eric Pin has selected Professor Éva Tardos (Cornell University, USA; http://www.cs.cornell.edu/~eva/) as the recipient of the EATCS Award 2017. Eva is the first female theoretical computer scientist who receives the EATCS Award.

The EATCS Award is given to acknowledge extensive and widely recognized contributions to theoretical computer science over a life-long scientific career. The list of the previous recipients of the EATCS Award is available at http://eatcs.org/index.php/eatcs-award.

The EATCS Award carries a prize money of 1000 Euros and will be presented at ICALP 2017, which will take place in Warsaw (Poland) from the 10th till the 14th of July 2017.

Sunday, January 22, 2017

Report on the rest of SOFSEM 2017 (guest post by Ignacio Fábregas)

I continue my notes on this year's SOFSEM conference where I left them. On Wednesday, I attended the tutorial session by Andrew Butterfield on Unifying Theories of Programming (UTP), a paradigm introduced by C.A.R. Hoare. Broadly speaking, UTP combines denotational, operational and algebraic semantics in a unified framework for both the formal specification and implementation. It is almost as there was no distinction between syntax and semantics. Andrew Butterfield also showed us  his tool UTP2, programmed in Haskell, and how the library for UTP in the Isabelle theorem prover works. 

One important part in any conference is to be able to discuss with other colleagues and experts about interesting papers. On Wednesday I missed the keynote talk by Jakko Hollmen about Predictive Analytics because I was talking about some papers with colleagues. (I found that to be a valid excuse to miss a talk in a conference!) Also, on Wednesday we had the social programme: an excursion with a guide around the city centre of Limerick and the conference dinner in the Bunratty Castle. Inside the castle we not only enjoyed a typical Irish dinner but also an entertainment program. It was a dinner with a medieval flavour because of the place but also because of the show we saw and the way we ate. We had a broth, pork ribs and chicken... But without forks! So we ate with our hands while listening to some typical Irish songs performed by the very talented people of the Bunratty Castle.

That's not all that happened on Wednesday. The organization had a surprise for us: since the lunch break was two hours long, during the second hour we had a talk about music and computers. Inside the Department of Computer Science in Limerick there is the Digital Media & Arts Research Centre, and the organization took advantage of this fact in order to propose them to give 3 talks: one on Wednesday and two on Thursday. The first talk was done by Mikael Fernström (http://www.idc.ul.ie/mikael/), who explained us his work. He uses computers in order to produce new sounds (frequencies that a physical instrument wouldn't be able to make) and algorithms in order to create music for "things". As an example, he created a piece that took some variables from meteorological data so we could listen to how the weather in Ireland sounds.

On Thursday, in this special sessions, Giuseppe Torre (http://www.muresearchlab.com/) showed us his work, where he uses computers to represent art. For example, his work 'AI PRISON' is a program in C++ that outputs '0' unless the technological singularity in Kurzweil's sense (or 'Terminator' sense) appears. It is exposed in the BLITZ Contemporary Art Gallery in Malta. Finally, Kerry l. Hagan (http://www.kerrylhagan.net/) played some of his musical works. Again, the use of computers is the cornerstone. She uses a program that allows her to program and compose the music she is going to play. This three artistic sessions where a really nice touch from the organization.

Back to the science, on Thursday we started with a keynote talk by Óscar Pastor López about Model-driven Development. The idea is that developers should be freed from programming concerns and be able to concentrate on guaranteeing that the final software product corresponds to what the company asked for. There are two basic notions in this field: capability and the Model Driven Architecture. The first one is especially used at the earliest steps of a software process in order to characterize the relevant modelling components to be specified. Whereas the Model Driven Architecture is used in the definition of the desired software process in order to structure the development from requirements to code.

Afterwards I attended the tutorial by Cristina Seceleanu about verification and test-case generation for Automotive Systems. Nowadays a car is developed with more electric components than hydraulic ones. For example, the connection between the steering wheel and the steering control is done by wires and software. This makes mandatory the use of testing in order to verify the correct behaviour of a car. In this field of research tools as UPPAAL are of great use. After the lunch we had the second keynote talk of the day. Paul Spirakis explained us about "programmable matter", which at the beginning seemed a field closest to sci-fi than actual computer science. So, let me explain what I understood, nowadays digital information (as internet) has spread along the world. That information is simply data, which can be manipulated. This vision is focused in our ability to manipulate data, and thus extended to manipulate matter via information-theoretic and computing mechanisms, incorporating information to the physical world. This is inspired in the biological world where living organisms are built by cells and those cells are built by information (DNA). Paul Spirakis explained us how Network Constructions is a promising technique for modelling this idea of "programmable matter". Broadly speaking, Network Constructions is simply a collection of finite-automata that interact randomly like molecules according to the rules of a common protocol.

The last session on Thursday for me was the tutorial session by Louis-Marie Traonouez by Plasma Lab, a statistical model checker (SMC). A SMC and uses less resources than a model checker at the cost of losing certainty. SMCs answers with a confidence level. Plasma Lab is SMC a library written in Java that can be used in tools as UPPAAL. The properties can be described in a form of bounded linear temporal logic (like LTL but with bounds in the length of the paths).

The last keynote talk of SOFSEM 2017 was on Friday and by Igor Walukiewicz. The topic of the talk was automatic verification of recursive programs with thread creation. Although this situation appears in programming languages such as Java, Scala or Erlang it doesn't behave well in automatic verification. For example, reachability is not decidable even for pushdown systems where two threads are communicating over a 2-bit shared variable. Igor Walukiewicz showed us some decidable setting by relaxing the semantics of thread creation operation. The obtaining result might not be very efficient but at least is computable.

Finally, I want to end these notes on SOFSEM by acknowledging the excellent work of the local organization at Limerick. It's true that SOFSEM 2017 didn't start very well in June, but all things considered everything went smooth and fine in Limerick. Also, the organization always tried to add something new to the conference (as those special talks about art and computer science), which is always nice for the participants.

Next year SOFSEM will be held in Austria and my recommendation is to check it out!

Wednesday, January 18, 2017

Report on the first two days of SOFSEM 2017 (guest post by Ignacio Fábregas)

Ignacio Fábregas has kindly sent me this report on the first two days of SOFSEM 2017, which is being held this week in Limerick, Ireland. I hope you'll enjoy reading it.  Many thanks to Ignacio for agreeing to write this guest post.
 
SOFSEM has always be an atypical conference with a format closest to a Winter School. This year marks a new beginning for the SOFSEM conference, since it's the first time it's held outside the Czech Republic or the Slovak Republic. As Jan van Leeuwen told us in the Welcome Reception on Monday, the goal of the Steering Committee of SOFSEM is to keep the spirit of SOFSEM alive while moving to a more typical format for a computer science conference. The main ingredients are still there: the keynote talks of renowned experts, tutorial sessions and the Student Research Forum; the only main difference is the country (Ireland) and the schedule (instead of having all the keynote talks the first and last day, now they are distributed across the conference days).

The first day of the Conference the keynote speaker was Kim G. Larsen. He told us about Cyber-Physical Systems (CPS), that is, systems combining computing elements with dedicated hardware and software having to monitor and control a particular physical environment. In order to study them, Larsen and his team propose a model-based approach, powered by the use of the tool Uppaal. An example of CPS is the adaptive cruise control for cars, an application where Europe is trying to match U.S. (where the Google Self-Driving car has already been approved by the legislation in several U.S. states) and where the team of Kim Larsen is making advances.

After the keynote talk we had the first parallel sessions: two sessions of the Foundations of Computer Science (FOCS) track and the first tutorial. I went to the FOCS session about Semantics, Specification and Compositionality, mainly because my talk was there :) But, more importantly, there was a talk of Uli Fahrenberg and another by Linda Bordo. Uli talked about behavioural specification theories for most equivalences in the linear-time–branching-time spectrum. He humbly defined his work as "almost" a rewriting of some old and overlooked results. Going back, reading and, most importantly, understanding, old results is a non-trivial exercise that computer science researchers should do more often. On the other hand, Linda talked about link-calculus, as a model that extends the point-to-point communication discipline of Milner’s CCS with multiparty interactions. She used links to build chains describing the flow of information among the agents participating in that interaction. The difficult part comes when deciding both the number of participants in an interaction and how they synchronize.

The second day of conference we had two keynote talks. Axel Legay talked about Software Product Lines (SPL), that is, the families of similar software products built from a common set of features. For example, like the mobile phones that a company as Samsung or HTC produce. He showed us how he and his team formalizes SPLs by means of what they called Featured Transition Systems and how the designed verification algorithms and tools to check temporal properties over them. The second keynote talk was by Marjan Mernik. He told us about Domain-Specific Languages that assist software developers in programming, and some open problems in the field like the lacking of tool support for different Domain-Specific Languages and the difficulties combining them.

Relative to the conference papers, in the FOCS session in Verification and Automated System Analysis we have talks by Mohammad Reza Mousavi, Nataliya Gribovskaya, Zhaowei Xu and Saket Saurabh. I want to highlight Mohammad's talk about the complexity of deriving invertible sequences from finite-state machines. The problem is the following: checking the existence of input/output sequences (UIOs) that identify states of the finite state machine specification (as it's the case of ioco) is PSPACE-complete; so some UIO generation algorithms utilise what are called invertible sequences; these allow one to construct additional UIOs once a UIO has been found. Mohammad and his coauthors studied some optimization problems and their complexity.

And that's all for now. Tomorrow (18 January 2017) we'll have fewer talks since we have the excursion and conference dinner. Also I'm planning to go for one tutorial session by Andrew Butterfield. I'll tell you more on that.

Thursday, January 12, 2017

Has the Feder-Vardi dichotomy conjecture been proved?

The Feder-Vardi dichotomy conjecture (which, as far as I know, stems from this 1998 article) is a celebrated open problem in the study of the complexity of constraint satisfaction problems (CSPs). It states that, for every finite relational structure, the set of all CSPs that can be represented using only relations chosen from that set is either in P or is NP-complete.

A paper claiming a solution to the Feder-Vardi  conjecture appeared yesterday on the arXiv:

https://arxiv.org/abs/1701.02409

Tómas Feder is the third author. (However, I noticed that the order of the authors in the version of the paper Feder posted on his web page is different from the one used in the arXiv version.)

The authors have also posted two videos on YouTube:

https://youtu.be/PBX51Qv5wtw
https://youtu.be/vD68km24Rh0

If the technical content of the paper is found to be correct by the community working on CSPs after careful peer review, this is a truly major result.

Wednesday, January 11, 2017

Call for Nominations for Scientific Directorship in Algorithms and Complexity at the Max Planck Institute for Informatics

The Max Planck Institute for Informatics in Saarbrücken is seeking nominations for the position of Scientific Member of the Max Planck Society and Director of the Institute. The founding director, Kurt Mehlhorn, will be retiring in the next three years, and the Institute is looking for a new scientific director for the algorithms and complexity department. (More precisely, Kurt will step down as Director on the day a new director joins. He will stay on as a researcher at the Institute.)

Webpage with the details of the call for nominations: http://www.mpi-inf.mpg.de/fileadmin/inf/d1/download/MPI-INF-Call-for-Nomations.pdf (Self-nominations are possible.)

Email: nominations@mpi-inf.mpg.de