Sensorimotor representation and knowledge-based reasoning for spatial exploration and localisation

We investigate a hybrid system for autonomous exploration and navigation, and implement it in a virtual mobile agent, which operates in virtual spatial environments. The system is based on several distinguishing properties. The representation is not map-like, but based on sensorimotor features, i.e. on combinations of sensory features and motor actions. The system has a hybrid architecture, which integrates a bottom-up processing of sensorimotor features with a top-down, knowledge-based reasoning strategy. This strategy selects the optimal motor action in each step according to the principle of maximum information gain. Two sensorimotor levels with different behavioural granularity are implemented, a macro-level, which controls the movements of the agent in space, and a micro-level, which controls its eye movements. At each level, the same type of hybrid architecture and the same principle of information gain are used for sensorimotor control. The localisation performance of the system is tested with large sets of virtual rooms containing different mixtures of unique and non-unique objects. The results demonstrate that the system efficiently performs those exploratory motor actions that yield a maximum amount of information about the current environment. Localisation is typically achieved within a few steps. Furthermore, the computational complexity of the underlying computations is limited, and the system is robust with respect to minor variations in the spatial environments.     

On the anatomy of social engineering attacks—A literature‐based dissection of successful attacks

The aim of this study was to explore the extent to which persuasion principles are used in successful social engineering attacks. Seventy‐four scenarios were extracted from 4 books on social engineering (written by social engineers) and analysed. Each scenario was split into attack steps, containing single interactions between offender and target. For each attack step, persuasion principles were identified. The main findings are that (a) persuasion principles are often used in social engineering attacks, (b) authority (1 of the 6 persuasion principles) is used considerably more often than others, and (c) single‐principle attack steps occur more often than multiple‐principle ones. The social engineers identified in the scenarios more often used persuasion principles compared to other social influences. The scenario analysis illustrates how to exploit the human element in security. The findings support the view that security mechanisms should include not only technical but also social countermeasures.     

Security-by-Experiment: Lessons from Responsible Deployment in Cyberspace

Conceiving new technologies as social experiments is a means to discuss responsible deployment of technologies that may have unknown and potentially harmful side-effects. Thus far, the uncertain outcomes addressed in the paradigm of new technologies as social experiments have been mostly safety-related, meaning that potential harm is caused by the design plus accidental events in the environment. In some domains, such as cyberspace, adversarial agents (attackers) may be at least as important when it comes to undesirable effects of deployed technologies. In such cases, conditions for responsible experimentation may need to be implemented differently, as attackers behave strategically rather than probabilistically. In this contribution, we outline how adversarial aspects are already taken into account in technology deployment in the field of cyber security, and what the paradigm of new technologies as social experiments can learn from this. In particular, we show the importance of adversarial roles in social experiments with new technologies.     

First-order Expressivity for S5-models: Modal vs. Two-sorted Languages

Standard models for model predicate logic consist of a Kripke frame whose worlds come equipped with relational structures. Both modal and two-sorted predicate logic are natural languages for speaking about such models. In this paper we compare their expressivity. We determine a fragment of the two-sorted language for which the modal language is expressively complete on S5-models. Decidable criteria for modal definability are presented.     

All Finitely Axiomatizable Tense Logics of Linear Time Flows Are CoNP-complete

We prove that all finitely axiomatizable tense logics with temporal operators for ‘always in the future’ and ‘always in the past’ and determined by linear fows time are coNP-complete. It follows, for example, that all tense logics containing a density axiom of the form ■ⁿ⁺¹_F p → ⁿ_F p, for some n ≥ 0, are coNP-complete. Additionally, we prove coNP-completeness of all ∩-irreducible tense logics. As these classes of tense logics contain many Kripke incomplete bimodal logics, we obtain many natural examples of Kripke incomplete normal bimodal logics which are nevertheless coNP-complete.     

Analysis of memory modulation by conditioned stimuli

Conditioned stimuli (CS) have multiple psychological functions that can potentially contribute to their effect on memory formation. It is generally believed that CS-induced memory modulation is primarily due to conditioned emotional responses, however, well-learned CSs not only generate the appropriate behavioral and physiological reactions required to best respond to an upcoming unconditioned stimulus (US), but they also serve as signals that the US is about to occur. Therefore, it is possible that CSs can impact memory consolidation even when their ability to elicit conditioned emotional arousal is significantly reduced. To test this, male Sprague–Dawley rats trained on a signaled active avoidance task were divided into “Avoider” and “Non-Avoider” subgroups on the basis of percentage avoidance after 6 d of training. Subgroup differences in responding to the CS complex were maintained during a test carried out in the absence of the US. Moreover, the subgroups displayed significant differences in stress-induced analgesia (hot-plate test) immediately after this test, suggesting significant subgroup differences in conditioned emotionality. Importantly, using the spontaneous object recognition task, it was found that immediate post-sample exposure to the avoidance CS complex had a similar enhancing effect on object memory in the two subgroups. Therefore, to our knowledge, this is the first study to demonstrate that a significant conditioned emotional response is not necessary for the action of a predictive CS on modulation of memory consolidation.

Biologically significant stimuli (unconditioned stimuli [US]) support learning and promote changes in behavior by enhancing the consolidation of memory (White and Milner 1992). Thus, stimuli such as food (Huston et al. 1974, 1977), pain (Galvez et al. 1996; Quirarte et al. 1998), and various drugs of abuse (Krivanek and McGaugh 1969; White 1996; Leri et al. 2013; Rkieh et al. 2014; Wolter et al. 2019, 2020) increase memory storage and facilitate performance on a variety of learning tasks when delivered during a window of memory consolidation that occurs following a learning experience (McGaugh and Roozendaal 2009; Roozendaal and Mcgaugh 2012; McGaugh 2015).Interestingly, exposure to stimuli paired with both incentive (Holahan and White 2013; Wolter et al. 2019, 2020; Baidoo et al. 2020) and aversive (Holahan and White 2002, 2004; Leong et al. 2015; Goode et al. 2016) USs also enhances memory consolidation, presumably because of the conditioned emotional responses that they generate. For example, CSs that precede exposure to footshock elicit freezing (Díaz-Mataix et al. 2017), avoidance (Dombrowski et al. 2013), analgesia (McNally et al. 1999), as well as sympathetic stimulation such as increases in heart rate (Zhang et al. 2019), blood pressure (Hsu et al. 2012), and release of stress hormones (Feenstra et al. 1999), all reactions that are elicited by footshock itself (Lim et al. 1982; McCarty and Baucom 1982; Conti et al. 1990; Galvez et al. 1996; O''Doherty 2004; Lázaro-Muñoz et al. 2010). Holahan and White (2002, 2004) reported that the memory enhancing action of a shock-paired CS could be blocked by lesions of the central amygdala nucleus (CeA), a region involved in generating the behavioral and neurohormonal responses to emotionally arousing stimuli (LeDoux 2003). As well, similarly to a range of aversive USs (anxiogenic drugs, predator odor, tail shock, restraint stress; Kim et al. 2001; Elliott and Packard 2008; Leong and Packard 2014), the effect of a CS paired with footshock on consolidation was found dependent on noradrenergic activation of the amygdala (Goode et al. 2016).However, well-learned CSs not only generate the appropriate behavioral and physiological reactions required to best respond to an upcoming US, but they also serve as signals that the US is about to occur. Temporal relationships between CSs and USs are learned rapidly during conditioning (Ohyama and Mauk 2001; Balsam et al. 2002), and these expectations modulate the expression of learned responses (Holland 2000; Balsam et al. 2010). Moreover, the ability of CSs to predict USs is heavily dependent on mesolimbic dopamine (DA) activity (Schultz et al. 1997; Flagel et al. 2011), and there is substantial evidence that mid-brain DA plays an important role in memory consolidation (White 1989; Managò et al. 2009; Redondo and Morris 2011; Yamasaki and Takeuchi 2017).This analysis suggests that CSs can impact memory consolidation because of their predictive function, even when their ability to elicit preparatory conditioned emotionality is significantly reduced. To test this idea, the current study used a signaled active avoidance task whereby rats learn to avoid an aversive US (footshock) by crossing from one compartment of a shuttle box to another during the presentation of a warning signal. Miller (1948) posited that animals perform the shuttle response during the signal because it prevents the occurrence of the US, and this reduces the experience of conditioned fear caused by the signal. However, it has been found that avoidance persists even when the warning signal no longer elicits a measurable fear state (Kamin et al. 1963; Linden 1969; Coover and Ursin 1973; Starr and Mineka 1977; Mineka and Gino 1980), suggesting that CSs can promote robust avoidance even though conditioned emotional responses are greatly reduced. The current study also used active avoidance because it consistently reveals robust individual differences in learning (Choi et al. 2010; Lázaro-Muñoz et al. 2010; Martinez et al. 2013; Antunes et al. 2020), such that animals can be distinguished into subgroups of Avoiders and Non-avoiders by simple median split (Storace et al. 2019) on percentage avoided USs. Although the source of the individual differences is unknown, it has been postulated the subgroups learn different behavioral responses to the CS (Lázaro-Muñoz et al. 2010; Martinez et al. 2013; Antunes et al. 2020): Avoiders display a loss of fear responses to the CS as the avoidance response is acquired (LeDoux et al. 2017; Cain 2018), while those who fail to acquire the avoidance response continue to display conditioned fear characterized by freezing (Martinez et al. 2013).By capitalizing on these individual differences, the current study explored whether both the predictive and preparatory functions of aversive CSs play a role in modulating consolidation of object memory using the spontaneous object recognition (OR) task. OR relies on the natural tendency of rats to explore novel objects (Winters et al. 2008) and this task was selected because it has been found sensitive to enhancement by exposure to contextual CSs paired with both incentive and aversive stimuli (Wolter et al. 2019, 2020; Baidoo et al. 2020). Given the evidence reviewed above, it was predicted that exposure to the avoidance CS complex (the training chamber, the retractable gate, the warning tone, and the cue light) would impact consolidation of object memory equally in Avoider and Non-Avoider subgroups. Avoider and Non-Avoider subgroups were tested for reactivity to thermal pain throughout avoidance training and testing using the hot-plate to provide an indirect measure of emotional reactivity to the footshock and/or to the aversive CS complex (Fig. 1). This approach was selected because fear/stress-inducing stimuli such as footshock (Maier and Watkins 1991; Rosellini et al. 1994), predator odor (Williams et al. 2005), and their CSs (Hotsenpiller and Williams 1997; McNally and Akil 2001; Ford et al. 2011), elicit stress-induced analgesia; a well-known defensive response in various species (Bolles and Fanselow 1980; Fendt and Fanselow 1999).Open in a separate windowFigure 1.Experimental design used in Experiments 1 and 2.