Lilia Georgieva

Hyperresolution for guarded formulae

This paper investigates the use of hyperresolution as a decision procedure and model builder for guarded formulae. In general, hyperresolution is not a decision procedure for the entire guarded fragment. However we show that there are natural fragments of the guarded fragment which can be decided by hyperresolution. In particular, we prove decidability of hyperresolution with or without splitting for the fragment GF1- and point out several ways of extending this fragment without losing decidability. As hyperresolution is closely related to various tableaux methods the present work is also relevant for tableaux methods. We compare our approach to hypertableaux, and mention the relationship to other clausal classes which are decidable by hyperresolution.

A New Clausal Class Decidable by Hyperresolution

In this paper we define a new clausal class, called \( \mathcal{B}\mathcal{U} \), which can be decided by hyperresolution with splitting. We also consider the model generation problem for \( \mathcal{B}\mathcal{U} \) and show that hyperresolution plus splitting can also be used as a Herbrand model generation procedure for \( \mathcal{B}\mathcal{U} \) and, furthermore, that the addition of a local minimality test allows us to generate only minimal Herbrand models for clause sets in \( \mathcal{B}\mathcal{U} \). In addition, we investigate the relationship of \( \mathcal{B}\mathcal{U} \) to other solvable classes.

Computational Space Efficiency and Minimal Model Generation for Guarded Formulae

This paper describes a number of hyperresolution-based decision procedures for a subfragment of the guarded fragment. We first present a polynomial space decision procedure of optimal worst-case space and time complexity for the fragment under consideration. We then consider minimal model generation procedures which construct all and only minimal Herbrand models for guarded formulae. These procedures are based on hyperresolution, (complement) splitting and either model constraint propagation or local minimality tests. All the procedures have concrete application domains and are relevant for multi-modal and description logics that can be embedded into the considered fragment.

Description logics for shape analysis

Verification of programs requires reasoning about sets of program states. In case of programs manipulating pointers, program states are pointer graphs. Verification of such programs involves reasoning about unbounded sets of graphs. Three-valued shape analysis (Sagiv et. al.) is an approach based on explicit manipulation of 3-valued shape graphs, which abstract sets of pointer graphs. Other approaches use symbolic representations, e.g., by describing (sets of) graphs as logical formulas. Unfortunately, many resulting logics are either undecidable or cannot express crucial properties like reachability and separation. In this paper, we investigate an alternative approach. We study well-known description logics as a framework for symbolic shape analysis. We propose a predicate abstraction based shape analysis, parameterized by description logics to represent the abstraction predicates. Depending on the particular logic chosen sharing, reachability and separation in pointer data structures are expressible.

Bounded model checking of pointer programs

We propose a bounded model checking procedure for programs manipulating dynamically allocated pointer structures. Our procedure checks whether a program execution of length n ends in an error (e.g. a NULL dereference) by testing if the weakest precondition of the error condition together with the initial condition of the program (e.g. program variable x points to a circular list) is satisfiable. We express error conditions as formulas in the 2-variable fragment of the Bernays-Schonfinkel class with equality. We show that this fragment is closed under computing weakest preconditions. We express the initial conditions by unary relations which are defined by monadic Datalog programs. Our main contribution is a small model theorem for the 2-variable fragment of the Bernays-Schonfinkel class extended with least fixed points expressible by certain monadic Datalog programs. The decidability of this extension of first-order logic gives us a bounded model checking procedure for programs manipulating dynamically allocated pointer structures. In contrast to SAT-based bounded model checking, we do not bound the size of the heap a priori, but allow for pointer structures of arbitrary size. Thus, we are doing bounded model checking of infinite state transition systems.

Anomaly Detection Using Agglomerative Hierarchical Clustering Algorithm

Intrusion detection is becoming a hot topic of research for the information security people. There are mainly two classes of intrusion detection techniques namely anomaly detection techniques and signature recognition techniques. Anomaly detection techniques are gaining popularity among the researchers and new techniques and algorithms are developing every day. However, no techniques have been found to be absolutely perfect. Clustering is an important data mining techniques used to find patterns and data distribution in the datasets. It is primarily used to identify the dense and sparse regions in the datasets. The sparse regions were often considered as outliers. There are several clustering algorithms developed till today namely K-means, K-medoids, CLARA, CLARANS, DBSCAN, ROCK, BIRCH, CACTUS etc. Clustering techniques have been successfully used for the detection of anomaly in the datasets. The techniques were found to be useful in the design of a couple of anomaly based Intrusion Detection Systems (IDS). But most of the clustering techniques used for these purpose have taken partitioning approach. In this article, we propose a different clustering algorithm for the anomaly detection on network datasets. Our algorithm is an agglomerative hierarchical clustering algorithm which discovers outliers on the hybrid dataset with numeric and categorical attributes. For this purpose, we define a suitable similarity measure on both numeric and categorical attributes available on any network datasets.

Issues in gender diversity and equality in the UK

There has been a significant increase in the number of initiatives to raise awareness of diversity-related challenges in technology worldwide within the past decade. Multiple organizations now emphasize a need for a close to 50%-50% male to female workforce distribution. Example of proposed activities include introducing quotas for women on board positions, promoting equal opportunities for employment in STEM (Science, Technology, Engineering, Mathematics) jobs and creating a woman-friendly work environment. However, despite these efforts, the growth of number of women working in STEM is still slow. To understand the impact of various initiatives and how they influence the work environment in universities in the UK, we conducted a survey to record responses from multiple women groups, so that we can identify the issues that they have been facing. This paper presents the insights drawn from the survey, along with recommendations for STEM and computing fields in order to increase female numbers in their programs. The survey presents qualitative measures of initiatives addressing the gender gap in the UK. The results show a clear need for prominent role models, mentoring, and promoting engagement of women in STEM subjects from an early age.

Formalisation and Verification of Knowledge Management in Digital Economy and Organisations

Knowledge management is the process of gathering, refining, organising, and disseminating knowledge through which organisations generate value from their intellectual and knowledge-based assets.

Modelling of Secure Data Transmission over a Multichannel Wireless Network in Alloy

This paper proposes a modelling and verification approach for data transmission over a multichannel wireless local area network (WLAN). The approach uses typed first-order logic as a specification language. We analyse a system which transmits data securely in the presence of the classic Man in The Middle (MitM) attack using Alloy. We develop a methodology for representing secure message exchange on a multi-channel WLAN which uses a changeable array and indices, instead of the message itself so that we can avoid both passive and active MitM attacks. We analyse the model for vulnerabilities and specify assertions for secure data transmission over a multichannel WLAN.

Modelling of security properties in Alloy

We study the problem of verification of security properties of Session Initiation Protocol (SIP) using the model analyser Alloy. We propose a novel approach to model analysis and demonstrating robustness of protocol models in first-order logic.