Helsinki Institute for Information Technology

About: Helsinki Institute for Information Technology is a facility organization based out in Espoo, Finland. It is known for research contribution in the topics: Population & Bayesian network. The organization has 630 authors who have published 1962 publications receiving 63426 citations.

Operational State Complexity under Parikh Equivalence

Abstract: We investigate, under Parikh equivalence, the state complexity of some language operations which preserve regularity. For union, concatenation, Kleene star, complement, intersection, shuffle, and reversal, we obtain a polynomial state complexity over any fixed alphabet, in contrast to the intrinsic exponential state complexity of some of these operations in the classical version. For projection we prove a superpolynomial state complexity, which is lower than the exponential one of the corresponding classical operation. We also prove that for each two deterministic automata A and B it is possible to obtain a deterministic automaton with a polynomial number of states whose accepted language has as Parikh image the intersection of the Parikh images of the languages accepted by A and B. Finally, we prove that for each finite set there exists a small context-free grammar defining a language with the same Parikh image.

Proceedings of SAT Challenge 2012 : Solver and Benchmark Descriptions

Abstract: This paper briefly describes the BossLS SAT Solver by explaining its basic functionality and outlining its features. I. THE k-SAT PROBLEM In the following, let F be a Boolean formula in conjunctive normal form (CNF), containing the n Boolean variables V = {x1, . . . , xn} in the form of 2n literals L = {x1,¬x1, . . . , xn,¬xn}. A formula in CNF is a conjunction (and) of disjunctions (or) of literals. The disjunctions of literals are called clauses. An assignment α : V → {0, 1} satisfies a formula in CNF, if and only if it assigns values to the variables such that in each clause there is at least one literal evaluating to true. Then, all clauses evaluate to true, and the formula’s conjunction evaluates to true as well (α(F ) = 1). Such assignments are called satisfying assignments or solutions (for F ). We call F satisfiable, if and only if there is at least one satisfying for it. SAT is a language that consists of all satisfiable Boolean formulas. The restriction to formulas in CNF, where a clause has at most k literals, is called k-SAT. The k-SAT problem is the problem to decide whether a given formula F is in k-SAT, that is, decide if the given formula is satisfiable. II. GENERAL IDEA BEHIND BOSSLS A SAT Solver is an algorithm that might solve the (k-)SAT problem for a given (CNF-)formula F . We call such a SAT Solver complete, if it can decide F ∈ SAT and F 6∈ SAT, and incomplete if it can decide F ∈ SAT but not F 6∈ SAT. The BossLS SAT solver is an incomplete SAT solver, following the approach of stochastic local search (SLS). Its basic functionality is equal to solvers like Sparrow [1] or gNovelty+ [8]. The general approach of these solvers is to assume that F is satisfiable by some (randomly created) assignment, and perform search by making local modifications to the assignment in order to increase the number of satisfied clauses. If a solution is found, it is used as a proof to support the claim that F is indeed satisfiable. If F is not satisfiable, the algorithm will not terminate. To be more precise: given F with n variables, SLS solvers will first create a random starting assignment α. They then check which of the clauses from F are not satisfied under α, that is α(Ci) = 0. Let this set of clauses be U . If U = ∅ then all clauses are evaluating to true, and then α is a solution. In this case, the search is over. If U 6= ∅, α can be no solution. SLS solvers then typically pick one of the clauses Ci ∈ U at random and try to fix the assignment α, such that α(Ci) = 1. This is done by selecting exactly one of the literals in α in order to invert its assignment. After flipping the assignment to the corresponding variable, and thereby modifying α into α′, the solvers will again check if all clauses are satisfied under this modified α′. The major difference between the solvers named above is how exactly they pick the literal from Ci in order to flip the corresponding variable assignment. III. PICKING LITERALS FOR FLIPPING THE CORRESPONDING VARIABLE ASSIGNMENT Let Ci ∈ F be a clause and let α(Ci) = 0. Assume, that Ci has been selected for fixing as explained in the previous section. Let w.l.o.g. Ci = (l1 ∨ . . . ∨ lk). The BossLS solver will now investigate each li and count the number of clauses that would become unsatisfied if it would indeed flip the assignment to the corresponding variable. This is called the break count of the variable and is denoted with blj := b(lj , α, F ) = b(¬lj , α, F ). Using these break counts, and a parameter called cb (the break base parameter value), the solver computes a function f of values for each lj ∈ Ci: f(lj , α, F ) = (cb) −blj (1) The probability for each literal in the clause to be selected for flipping then follows by computing [2]: P (select lj ∈ Ci) = f(lj , α, F ) ∑ lw∈Ci f(lw, α, F ) . (2) According to this distribution, it picks a literal at random and then flips the corresponding variable assignment. According to [2], this scheme gives superior performance for random 3SAT formulae. We refer the reader to this paper for additional information. IV. HANDLING CRAFTED INSTANCES

On application of Host Identity Protocol in wireless sensor networks

Abstract: Recent advances in development of low-cost wireless sensor platforms open up opportunities for novel wireless sensor network (WSN) applications. Likewise emerge security concerns of WSNs receiving closer attention of research community. Well known security threats in WSNs range from Denial-of-Service (DoS), Replay and Sybil attacks to those targeted at violating data integrity and confidentiality. Public-key cryptography (PKC) as a countermeasure to potential attacks, although originally treated infeasible for resource-constrained sensor nodes, has shown its eligibility for WSNs in the past few years. However, different security and performance requirements, energy consumption issues, as well as varying hardware capabilities of sensor motes pose a challenge of finding the most efficient security protocol for a particular WSN application and scenario. In this paper, we propose to use the Host Identity Protocol (HIP) as the main component for building network-layer security in WSNs. Combining PKC signatures to authenticate wireless nodes, a Diffie-Hellman key exchange to create a pairwise secret key, a puzzle mechanism to protect against DoS attacks and the IPsec protocol for optional encryption of sensitive application data, HIP provides a standardized solution to many security problems of WSNs. We discuss how HIP can strengthen security of WSNs, suggest possible alternatives to its heavy components in particular WSN applications and evaluate their computational and energy costs on a Linux-based Imote2 wireless sensor platform.

Explaining mixture models through semantic pattern mining and banded matrix visualization

Abstract: This paper presents an approach to semi-automated data analysis, supported by tools for pattern construction, exploration and explanation. The proposed three-part methodology for multiresolution 0---1 data analysis consists of data clustering with mixture models, extraction of rules from clusters, as well as data and rule visualization using banded matrices. The results of the three-part process: clusters, rules from clusters, and banded structure of the data matrix are finally merged in a unified visual banded matrix display. The incorporation of multiresolution data is enabled by the supporting ontology, describing the relationships between the different resolutions, which is used as background knowledge in the semantic pattern mining process of descriptive rule induction. The presented experimental use case highlights the usefulness of the proposed methodology for analyzing complex DNA copy number amplification data, studied in previous research, for which we provide new insights in terms of induced semantic patterns and cluster/pattern visualization. The methodology is successfully evaluated on four other publicly available data sets, which further demonstrates the utility of the proposed approach.