In the rapidly evolving landscape of smartphones and handheld devices, Android malware stands as a substantial security concern. Employing static analysis for mobile malware presents a proactive approach to understanding and unveiling potential threats within Android applications without the need for execution. Most of the existing studies on static analysis rely on machine learning. However, the black-box nature of machine learning models and their lack of explainability often hinder trust, transparency, and the ability to understand or justify their predictions. Counterfactual explanations enable security analysts to grasp the reasoning behind the decisions of black-box machine learning models (the “why?”) and also offer a way to pinpoint specific data instances whose alteration would lead to different prediction results (the “why not?”). In this paper, we investigate the use of the counterfactual explanation method to explain the predictions made by a machine learning model for Android malware classification. We assessed the quality of counterfactual explanations using a stability metric as well as investigating their feasibility by defining feature-based constraints.

We introduce a software tool for reasoning about belief change in situations where information is received from reports and observations. Our focus is on the interaction between trust and belief. The software is based on a formal model where the level of trust in a reporting agent increases when they provide accurate reports and it decreases when they provide innaccurate reports. If trust in an agent drops below a given threshold, then their reports no longer impact our beliefs at all. The notion of accuracy is determined by comparing reports to observations, as well as to reports from more trustworthy agents. The emphasis of this paper is not on the formalism; the emphasis is on the development of the prototype system for automatically calculating the result of iterated revision problems involving trust. We present an implemented system that allows users to flexibly specify and solve complex revision problems involving reports from partially trusted sources.

Consider a set of agents with initial beliefs and a formal operator for incorporating new information. Now suppose that, for each agent, we have a formula that we would like them to believe. Does there exist a single announcement that will lead all agents to believe the corresponding formula? This paper studies the problem of the existence of such an announcement in the context of model-preference definable revision operators. First, we provide two characterisation theorems for the existence of announcements: one in the general case, the other for total partial orderings. Second, we exploit the characterisation theorems to provide upper bound complexity results. Finally, we also provide matching optimal lower bounds for the Dalal and Ginsberg operators.

Belief revision is concerned with incorporating new information into a pre-existing set of beliefs. When the new information comes from another agent, we must first determine if that agent should be trusted. In this paper, we define trust as a pre-processing step before revision. We emphasize that trust in an agent is often restricted to a particular domain of expertise. We demonstrate that this form of trust can be captured by associating a state partition with each agent, then relativizing all reports to this partition before revising. We position the resulting family of trust-sensitive revision operators within the class of selective revision operators of Ferme and Hansson, and we prove a representation result that characterizes the class of trust-sensitive revision operators in terms of a set of postulates. We also show that trust-sensitive revision is manipulable, in the sense that agents can sometimes have incentive to pass on misleading information.

Public announcements cause each agent in a group to modify their beliefs to incorporate some new piece of information, while simultaneously being aware that all other agents are doing the same. Given a set of agents and a set of epistemic goals, it is natural to ask if there is a single announcement that will make each agent believe the corresponding goal. This problem is known to be undecidable in a general modal setting, where the presence of nested beliefs can lead to complex dynamics. In this paper, we consider not necessarily truthful public announcements in the setting of AGM belief revision. We prove that announcement finding in this setting is not only decidable, but that it is simpler than the corresponding problem in the most simplified modal logics. We then describe AnnB, an implemented tool that uses announcement finding as the basis for controlling robot behaviour through belief manipulation.

Forensic examinations of a mobile phone that consider only the internal memory can miss potentially vital data that is accessible from the device, but not stored locally. In this paper, we look at a forensic tool that is able to download data stored on the cloud, using credentials gleaned from device extractions. Through experimention with a variety of devices and configurations, we examine the effectiveness of the software for its stated purpose. The results suggest that we are able to obtain information from the cloud in this manner, but only under some relatively strong assumptions. Practical issues and legal considerations are discussed.

We discuss two related forms of trust. One form of trust is related to the perceived knowledge of other agents; we accept the information that another agent provides if we believe they have sufficient expertise in a particular domain. The second form is related to action; we trust another agent to act on our behalf if we believe they will choose acceptable actions. In this paper, we explore the relationship between these two forms of trust. In particular, we use an existing model of trust to demonstrate how trust over knowledge can determine when trust over actions is appropriate. We take a formal approach to this problem, using logic-based tools for representing and reasoning about actions and beliefs to characterize trust over action. While our primary aim is to develop a formal methodology that permits trust over actions to be defined in terms of trust over knowledge, we also consider applications that are both practical and speculative. On the practical side, we consider how our methods can be used to reason about trusted third parties in communication protocols. On the speculative side, we suggest that models of trust have a role to play in the development of ethical decision-making agents.

We address security vulnerabilities for a smart thermostat. As this kind of smart appliance is adopted in homes around the world, every user will be opening up a new avenue for cyber attack. Since these devices have known vulnerabilities and they are being managed by non-technical users, we anticipate that smart thermostats are likely to be targetted by unsophisticated attackers relying on publicly available exploits to take advantage of weakly protected devices. As such, in this paper, we take the role of a `script kiddy' and we assess the security of a smart thermostat by using Internet resources for attacks at both the physical level and the network level. We demonstrate that such attacks are unlikely to be effective without some additional social engineering to obtain user credentials. Moreover, we suggest that the vulnerability to attack can be further minimized by simply reducing the use of remote storage where possible.

We consider the iterated belief change that occurs following an alternating sequence of actions and observations. At each instant, an agent has beliefs about the actions that have occurred as well as beliefs about the resulting state of the world. We represent such problems by a sequence of ranking functions, so an agent assigns a quantitative plausibility value to every action and every state at each point in time. The resulting formalism is able to represent fallible belief, erroneous perception, exogenous actions, and failed actions. We illustrate that our framework is a generalization of several existing approaches to belief change, and it appropriately captures the non-elementary interaction between belief update and belief revision.

In this paper, we explore the use of ranking functions in reasoning about belief change. It is well-known that the semantics of AGM belief revision can be defined either through total pre-orders or through ranking functions over interpretations. While both approaches have generally been seen as equivalent with respect to single-shot belief revision, we argue that ranking functions provide distinct advantages at both the theoretical level and the practical level. We demonstrate belief revision induces a natural algebra over ranking functions, which treats belief states and observations in the same manner. Moreover, when we introduce belief progression due to actions, we demonstrate that many natural domains can be easily represented with suitable ranking functions. We support our position through formal results, as well as a series of natural problems in commonsense reasoning. We conclude with a discussion of aggregate functions for combining rankings, as well as potential future applications in counterfactual reasoning.

In order to model the dynamically changing beliefs of an agent, one must actually address two distinct issues. First, one must devise a model of static beliefs that accurately captures the appropriate notions of incompleteness and uncertainty. Second, one must define appropriate operations to model the way beliefs are modified in response to different events. Historically, the former is addressed through the use of modal logics and the latter is addressed through belief change operators. However, these two formal approaches are not particularly complementary; the normal representation of belief in a modal logic is not suitable for revision using standard belief change operators. In this paper, we introduce a new modal logic that uses the accessibility relation to encode epistemic entrenchment, and we demonstrate that this logic captures AGM revision. We consider the suitability of our new representation of belief, and we discuss potential advantages to be exploited in future work.

In action domains where agents may have erroneous beliefs, reasoning about the effects of actions involves reasoning about belief change. In this paper, we use a transition system approach to reason about the evolution of an agent's beliefs as actions are executed. Some actions cause an agent to perform belief revision while others cause an agent to perform belief update, but the interaction between revision and update can be non-elementary. We present a set of rationality properties describing the interaction between revision and update, and we introduce a new class of belief change operators for reasoning about alternating sequences of revisions and updates. Our belief change operators can be characterized in terms of a natural shifting operation on total pre-orderings over interpretations. We compare our approach with related work on iterated belief change due to action, and we conclude with some directions for future research.

Cryptographic protocols are usually specified in an informal, ad hoc language, with crucial elements, such as the protocol goal, left implicit. We suggest that this is one reason that such protocols are difficult to analyse, and are subject to subtle and nonintuitive attacks. We present an approach for formalising and analysing cryptographic protocols in a theory of action, specifically the situation calculus. Our thesis is that all aspects of a protocol must be explicitly specified. We provide a declarative specification of underlying assumptions and capabilities in the situation calculus. A protocol is translated into a sequence of actions to be executed by the principals, and a successful attack is an executable plan by an intruder that compromises the specified goal. Our prototype verification software takes a protocol specification, translates it into a high-level situation calculus (Golog) program, and outputs any attacks that can be found. We describe the structure and operation of our prototype software, and discuss performance issues.

The spectrum of a first-order sentence is the set of cardinalities of its finite models. Relatively little is known about the subclasses of spectra that are obtained by looking only at sentences with a specific signature. In this paper, we study natural subclasses of spectra and their closure properties under simple subdiagonal functions. We show that many natural closure properties turn out to be equivalent to the collapse of potential spectrum hierarchies. We prove all of our results using explicit transformations on first-order structures.