Section: Highlights of the Year

Highlights of the Year

Containment for RDF Schemas

The ShEx language for defining RDF schemas was proposed and developed earlier by the Links team in cooperation with the W3C. Slawek Staworko et al. now studied the containment problem for ShEx schemas for RDF documents. They showed at PODS [7] – the best database theory conference – that the problem is decidable, but co-NEXP -hard. This is joined work with P. Wieczorek from the University of Wrazlaw.

Foundations of AI: Knowledge Compilation

Florent Capelli et al. showed at STACS [15] – a top conferences in theoretical computer science – a new knowledge compilation procedure for quantified boolean formulas allowing to decide the satisfiability quantified boolean formulas with bounded tree width in polynomial time. This can be applied in particular to first-order database queries with quantifiers. This is joined work with S. Mengel from the CNRS in Lens.

Foundations of AI: Constrained Topologial Sort

Charles Paperman et al. showed at ICALP [8] – a top conferences in theoretical computer science – how to compute efficiently topological sorts of graphs under regular constraints. The problem was initially introduced in the context of preferential query answer for uncertain databases, where on usually wants to sort the query answers by some preferences, that are known only partially. It becomes then crucial to look for total orders on the answer set satisfying regular constraints that specify the preferences. Finding such an order for regular constraints was know to be infeasible in general. In this article, a class of regular constraints is identified for which this problem becomes tractable. A (partial) decidable dichotomy theorem is proven drawing the frontier between the kind of constraints which are feasible from those which are not. This is joined work with A. Amarilli from Telecom Paristech.