favorite3Yet, in difference from that work, we apply our analysis to learners and to the attributes comprising the dataset they aim to learn from.
favorite8Validity of model assumptions In Section Model assumptions it was assumed that, given a dataset of features and its partitioning into Bi s, any choice of a feature from each Bi and a learning method that uses the chosen features results in comparable performance.
favorite5Thus, as a result of the attack the feature x will appear less desirable to the learner and, consequently, may not be chosen for the learning..
favorite2Tampering with learning attributes As stated earlier, in our model learners and adversaries select attributes.
favorite0There are cases in which an adversary can strategically tamper with the input data to affect the outcome of the learning process.
favorite8While the students encounter the concepts of state machines, abstraction and composition at other IS courses (such as modeling and design), aspects related to working with formal specifications are not covered elsewhere.
favorite12It discusses teaching formal methods at universities of applied sciences, where there are usually limiting factors which are relevant to the IS context as well: (i) students have very limited theoretical background, and (ii) they are strongly focused on the direct applicability of what they are taught.
favorite85Wing stresses the importance of integrating formal methods into the existing CS curriculum by teaching their common conceptual elements, including state machines, invariants, abstraction, composition, induction, specification and verification.
favorite10In what follows we briefly survey previous reflections on the content of logic and formal methods courses that practitioners really need and their integration into the curricula, and propose how to adapt the proposed ideas for the context of IS..
favorite6We report on our experience in designing and teaching the course "Logic and Formal Specification" to graduate students at the Information Systems (IS) department at the University of Haifa, which is one of the few to include a mandatory course on logic and formal methods in its graduate study program.
favorite18While in agile CTD an iterative process concerns only the interaction between a model and a test plan, in the current framework we have several inter-related levels of abstraction and their corresponding test plans.
favorite1A test plan is a triple P = (E, C, T ), where E is a combinatorial model, C is a set of interactions called coverage requirements, and T is a set of scenarios called tests, where T covers C.
favorite1Contributions: We propose a human-centered agile modelling and testing approach for cyber-physical systems, which combines two types of refinements: static (or system-oriented, meant to hide unnecessary details) and dynamic (or testeroriented, meant to provide the ability to correct and complete the developed artefacts).
favorite5More concretely, discovery of an error or incomplete information may cause the tester to return to the model and refine it, which in its turn may induce further changes in the existing test plan.
favorite17In practice, however, due to the complex and error prone character of modelling and test planning, the modeller/tester often makes mistakes and may revisit the different levels of abstractions to make dynamic refinements.
favorite6In this paper we present our ongoing work on introducing humancentred considerations into modelling and testing of CPSs, which allow for agile iterative refinement processes of different levels of abstraction when errors are discovered or missing information is completed.