This paper develops a novel mathematical framework for collaborative learning by means of geometrically inspired kernel machines which includes statements on the bounds of generalisation and approximation errors, and sample complexity. For classification problems, this approach allows us to learn bounded geometric structures around given data points and hence solve the global model learning problem in an efficient way by exploiting convexity properties of the related optimisation problem in a Reproducing Kernel Hilbert Space (RKHS). In this way, we can reduce classification problems to determining the closest bounded geometric structure from a given data point. Further advantages that come with our solution is that our approach does not require clients to perform multiple epochs of local optimisation using stochastic gradient descent, nor require rounds of communication between client/server for optimising the global model. We highlight that numerous experiments have shown that the proposed method is a competitive alternative to the state-of-the-art.

This paper develops a novel mathematical framework for collaborative learning by means of geometrically inspired kernel machines which includes statements on the bounds of generalisation and approximation errors, and sample complexity. For classification problems, this approach allows us to learn bounded geometric structures around given data points and hence solve the global model learning problem in an efficient way by exploiting convexity properties of the related optimisation problem in a Reproducing Kernel Hilbert Space (RKHS). In this way, we can reduce classification problems to determining the closest bounded geometric structure from a given data point. Further advantages that come with our solution is that our approach does not require clients to perform multiple epochs of local optimisation using stochastic gradient descent, nor require rounds of communication between client/server for optimising the global model. We highlight that numerous experiments have shown that the proposed method is a competitive alternative to the state-of-the-art.

Abstract This paper presents a novel simplification calculus for propositional logic derived from Peirce’s existential graphs’ rules of inference and implication graphs. Our rules can be applied to propositional logic formulae in nested form, are equivalence-preserving, guarantee a monotonically decreasing number of variables, clauses and literals, and maximise the preservation of structural problem information. Our techniques can also be seen as higher-level SAT preprocessing, and we show how one of our rules (TWSR) generalises and streamlines most of the known equivalence-preserving SAT preprocessing methods. In addition, we propose a simplification procedure based on the systematic application of two of our rules (EPR and TWSR) which is solver-agnostic and can be used to simplify large Boolean satisfiability problems and propositional formulae in arbitrary form, and we provide a formal analysis of its algorithmic complexity in terms of space and time. Finally, we show how our rules can be further extended with a novel n-ary implication graph to capture all known equivalence-preserving preprocessing procedures.

This article proposes a user-adaptable and personalized authentication paradigm for healthcare organizations, which anticipates to seamlessly reflect patients’ episodic and autobiographical memories to graphical and textual passwords aiming to improve the security strength of user-selected passwords and provide a positive user experience. We report on a longitudinal study that spanned over 3 years in which three public European healthcare organizations participated to design and evaluate the aforementioned paradigm. Three studies were conducted ( n = 169) with different stakeholders: (1) a verification study aiming to identify existing authentication practices of the three healthcare organizations with diverse stakeholders ( n = 9), (2) a patient-centric feasibility study during which users interacted with the proposed authentication system ( n = 68), and (3) a human guessing attack study focusing on vulnerabilities among people sharing common experiences within location-aware images used for graphical passwords ( n = 92). Results revealed that the suggested paradigm scored high with regard to users’ likeability, perceived security, usability, and trust, but more importantly it assists the creation of more secure passwords. On the downside, the suggested paradigm introduces password guessing vulnerabilities by individuals sharing common experiences with the end users. Findings are expected to scaffold the design of more patient-centric knowledge-based authentication mechanisms within today's dynamic computation realms.

DataMod 2020 aims to bring together practitioners and researchers from academia, industry and research institutions interested in the combined application of computational modelling methods with data-driven techniques from the areas of knowledge management, data mining and machine learning. Modelling methodologies of interest include automata, agents, Petri nets, process algebras and rewriting systems. Application domains include social systems, ecology, biology, medicine, smart cities, governance, security, education, software engineering, and any other field that deals with complex systems and large amounts of data. Papers can present research results in any of the themes of interest for the symposium as well as application experiences, tools and promising preliminary ideas. Papers dealing with synergistic approaches that integrate modelling and knowledge management/discovery or that exploit knowledge management/discovery to develop/syntesise system models are especially welcome.

The Second International Workshop on Adaptive and Personalized Privacy and Security (APPS 2020) aims to bring together researchers and practitioners working on diverse topics related to understanding and improving the usability of privacy and security software and systems, by applying user modeling, adaptation and personalization principles. Our special focus in 2020 is on healthcare systems, more specifically on ensuring security and privacy of medical data in smart patient-centric healthcare systems. The second edition of the workshop includes interdisciplinary contributions from Austria, Canada, China, Cyprus, Denmark, Germany, Greece, Israel, the Netherlands, Turkey and the UK that introduce new and disruptive ideas, suggest novel solutions, and present research results about various aspects (theory, applications, tools) for bringing user modeling, adaptation and personalization principles into privacy and systems security. This summary gives a brief overview of APPS 2020, held online in conjunction with the 28th ACM Conference on User Modeling, Adaptation and Personalization (ACM UMAP 2020).

Future-generation healthcare systems will be highly distributed, combining centralised hospital systems with decentralised home-, work-and environment-based monitoring and diagnostics systems. These will reduce costs and injury-related risks whilst both improving quality of service, and reducing the response time for diagnostics and treatments made available to patients. To make this vision possible, medical data must be accessed and shared over a variety of mediums including untrusted networks. In this paper, we present the design and initial implementation of the SERUMS tool-chain for accessing, storing, communicating and analysing highly confidential medical data in a safe, secure and privacy-preserving way. In addition, we describe a data fabrication framework for generating large volumes of synthetic but realistic data, that is used in the design and evaluation of the tool-chain. We demonstrate the present version of our technique on a use case derived from the Edinburgh Cancer Centre, NHS Lothian, where information about the effects of chemotherapy treatments on cancer patients is collected from different distributed databases, analysed and adapted to improve ongoing treatments.

. For patients suffering from two or more chronic conditions, known as multimorbidity, several guidelines have to be applied simultaneously, which may lead to severe adverse effects when the combined recommendations and prescribed medications are inconsistent or incomplete. This paper presents an automated formal framework to detect, highlight and resolve conflicts in the treatments used for patients with multimorbidities focusing on medications. The presented extended framework has a front-end which takes guidelines captured in a standard modelling language and returns the visualisation of the detected conflicts as well as suggested alternative treatments. Internally, the guidelines are transformed into formal models capturing the possible unfoldings of the guidelines. The back-end takes the formal models associated with multiple guidelines and checks their correctness with a theorem prover, and inherent inconsistencies with a constraint solver. Key to our approach is the use of an optimising constraint solver which enables us to search for the best solution that resolves/minimises conflicts according to medication efficacy and the degree of severity in case of harmful combinations, also taking into account their temporal overlapping. The approach is illustrated throughout with a real medical example.

Common chronic conditions are routinely treated following standardised procedures known as clinical pathways. For patients suffering from two or more chronic conditions, referred to as multimorbidities, several pathways have to be applied simultaneously. However, since pathways rarely consider the presence of comorbidities, applying several pathways may lead to potentially harmful (medication) conflicts. This paper proposes an automated framework to detect, highlight and resolve conflicts in the treatments used for patients with multimorbidites. We use BPMN as a modelling language for capturing care guidelines. A BPMN model is transformed into an intermediate formal model capturing the possible unfoldings of the pathway. Putting together the constraint solver Z3 and the theorem prover Isabelle, we combine treatment plans and check the correctness of the approach. We illustrate the approach with an example from the medical domain and discuss future work.

In complex system design, it is important to construct several design models focusing on different aspects of a system to gain a better understanding of individual component structure and behaviour. Scenarios of execution are commonly used to specify partial behaviour and interactions between a group of system objects or components. However, partial specifications may hide inconsistencies or an otherwise unintentionally incomplete or underspecified behavioural model. This paper proposes a new powerful technique combining constraint solvers and theorem provers to complete partial specifications and determine overall model inconsistencies. We use a true-concurrent model, namely labelled event structures, which can be used as the underlying semantics of widely used workflow or scenario-based languages. We show how an interplay between the theorem prover Isabelle and constraint solver Z3 can be used for detecting and solving partial specifications and inconsistencies over event structures.

Design of large systems often involves the creation of models that describe partial specifications. Model composition is the process of combining partial models to create a single coherent model. This paper presents an automatic composition technique for creating a sequence diagram from partial specifications captured in multiple sequence diagrams with the help of Alloy. Our contribution is twofold: a novel true-concurrent semantics for sequence diagram composition, and a model-driven transformation of sequence diagrams onto Alloy that preserves the semantics of composition defined. We have created a tool called SD2Alloy that implements the automated technique and works as follows: two given sequence diagrams are transformed in two Alloy models, and then combined according to a set of logical constraints, determined by our compositional semantics, into a final composed Alloy model. The technique can also be used to detect problems and inconsistencies in the composition of diagrams.

We present an approach for flexible analysis of complex system models based on a parametric model-to-model (M2M) transformation where the target model has variants. We describe a M2M transformation from sequence diagrams (SDs) to coloured Petri nets (CPNs) with (untimed, timed, stochastic) variants enabling different forms of dependability and performance analysis. The transformation is parametric on the chosen variant with the core set of rules defining the transformation from SDs to CPNs. Moreover, the flexibility of the framework lies in the incremental nature of the transformation: given a SD (with stochastic and time annotations) and corresponding untimed CPN, we can generate other CPN variants by incrementally applying the specific variant rules. This paper contributes towards the theoretical foundations of parametric transformations, defines and proves the semantic correctness of a parametric transformation between SDs and CPN variants.

In this paper we define a formal strongly consistent transformation from UML2 sequence diagrams to coloured Petri nets (CPNs). We define the language of sequence diagrams (set of legal traces) and show how this is mapped onto an equivalent language of CPNs through formal transformation rules. The rules defined can be extended to consider modelling aspects such as stochastic and real-time behaviour. This work is part of a more general framework based on model driven transformations to validate UML models using Petri nets, and hence exploit existing Petri net analysis and verification tools for UML-based design.

Frameworks are increasingly being recognised as very useful components in the emerging paradigm of component-based software development (CBD). They are widely accepted as better units of reuse than objects. The CBD methodology Catalysis, for instance, uses frameworks. However, at present, Catalysis frameworks are described only informally, which means we cannot reason formally about frameworks, in particular their composition, and thereby their reuse. For this, we would need to specify frameworks (and their composition) formally. In this paper, we describe our approach for doing so.

For purposes of security analysis, cryptographic protocols can be translated from a high-level message-list language such as CAPSL into a multiset rewriting (MSR) rule language such as CIL. The natural translation creates two rules per message or computational action. We show how to optimize the natural rule set by about 50% into a form similar to the result of hand encoding, and prove that the transformation is sound because it is attack-preserving, and unique because it is terminating and confluent. The optimization has been implemented in Java.