Mikołaj Bojańczyk
October 2, 2017

Hilbert’s Basis Theorem and transducers

These slides show how to decide equivalence of tree-to-string transducers using the Hilbert Basis Theorem. The idea is based on

Helmut Seidl, Sebastian Maneth, Gregor Kemper:
Equivalence of Deterministic Top-Down Tree-to-String Transducers is Decidable. FOCS 2015: 943-962

There are two parts:

  1. consider grammars that generate tuples of rational numbers, and use the Hilbert Basis Theorem to show that the following is decidable: “given a grammar, decide if the only tuple that it generates is the zero tuple”. slides
  2. reduce the equivalence problem for transducers to the zero-ness problem discussed above. slides
May 8, 2017

Lipa Summer School (July 3-6, 2017)

The Lipa Summer School (click here for the page of the school, including registration) is a summer school on topics connected to logic in computer science. There will be 4 mini-courses given by:

  • Stephan Kreutzer (Berlin) Algorithmic meta-theorems
  • Joël Ouaknine (Saarbrücken) Decision Problems for Linear Recurrence Sequences
  • Moshe Vardi (Rice)  Linear-time verification and synthesis
  • Mikołaj Bojańczyk (Warsaw, organiser) What is a recognisable language?

 

March 27, 2017

Positions (deadline April 30, 2017)

I am leading an ERC Consolidator grant on automata and logic called “A unified theory of finite-state recognisability”. The main goal is to find algebraic structures (like monoids) for things like infinite trees or profinite words. However, I can be very flexible on the topic, as long as it is tangentially connected to logic or automata.

I am searching for talented postdocs and phd students. The deadline for applications is April 30, 2017. Send applications to me by email. Please feel to ask questions before that date!

Postdocs. The postdoc can begin at any time from June 1 until the end of 2017. The duration is 6 – 12 months, with a possibility of extension for another year. The application consists of your name and 1-2 names of people who can provide references. I am looking for people with a good track record in the top conferences of the field. Teaching is optional.

PhD students. A typical duration is 4 years, starting any time from June 1 until the end of 2017. In application please include a cv and 1-2 names of people who can provide references. I am looking for candidates with a strong background in mathematics.

August 11, 2015

Tree width à la metro

Here is a way of drawing graphs of bounded tree width, inspired by metro maps. The horizontal black lines are graph edges, the coloured shapes are vertices, and the gray background is the tree underlying the decomposition.

tree-width-metro

 

The mathematical content of the picture is that a graph has tree width k if and only if it is a subgraph of a (k+1)-colourable chordal graph (thanks to Michał Pilipczuk for pointing this out).

June 20, 2015

EATCS council

In 2013, I joined the EATCS council. Here is what I promised to do:


 

I am a candidate for the EATCS council. My program consists of three items. The first item is something where I promise concrete action, the remaining two are more speculative.

  1. I will support open access, in particular for ICALP proceedings.
  2. In the long term, I think that the TCS community should move from publishing in conferences to publishing in journals. The ensuing travel is a waste of time and money. It is not clear to me how the EATCS can help solve this problem right now, without losing the accumulated prestige capital of ICALP. Maybe ICALP could move to a continuous submission model like VLDB?
  3. It is my impression that the current EATCS members are mainly people who have recently attended ICALP, which is a bit random. Membership should be made more representative. Maybe we should make membership free (but still feed EATCS from an ICALP tax)?

 

Update. The first item is solved: starting with 2016, ICALP will be using the open access LIPICS system. As it turned out, this idea had overwhelming support inside the council.