Project description:Appetitive aggression is the attraction to violent behavior, which can peak in the experience of a combat high. In various war and conflict scenarios, members of armed groups have reported developing a desire to hunt and even kill humans. More recently, we reported that the phenomenon has also been observed in female ex-combatants with varying participation in warfare. Despite recent investigations on risk factors for appetitive aggression, sex-specific pathways in the development of appetitive aggression have not yet been delineated. This study investigated moderation effects of sex on previously identified risk factors for appetitive aggression by means of regression analyses in a sample of individuals with varying degrees of warfare participation (overall sample, n = 602). First examining a sample characterized by backgrounds heterogeneous in both sociodemographic data and war experiences, the analysis was then replicated in a subsample of fighters active during the civil war (combatant sample, n = 109). In both samples, regression analyses revealed significant moderation effects of sex. Childhood maltreatment and traumatic events had positive associations on the development of appetitive aggression for males but a negative (childhood maltreatment) or no (traumatic events) association for females. Perpetrated events were more strongly correlated with appetitive aggression for females than for males. This pattern was pronounced for the combatant sample. These results are in favor of sex-linked pathways. In both sexes, appetitive aggression may have evolved as a biologically prepared response to cruel environments but might develop along different trajectories. The current study highlights the need for addressing appetitive aggression in order to support peace-building processes and emphasizes sex specific starting-points.
Project description:The automatic detection of violent actions in public places through video analysis is difficult because the employed Artificial Intelligence-based techniques often suffer from generalization problems. Indeed, these algorithms hinge on large quantities of annotated data and usually experience a drastic drop in performance when used in scenarios never seen during the supervised learning phase. In this paper, we introduce and publicly release the Bus Violence benchmark, the first large-scale collection of video clips for violence detection on public transport, where some actors simulated violent actions inside a moving bus in changing conditions, such as the background or light. Moreover, we conduct a performance analysis of several state-of-the-art video violence detectors pre-trained with general violence detection databases on this newly established use case. The achieved moderate performances reveal the difficulties in generalizing from these popular methods, indicating the need to have this new collection of labeled data, beneficial for specializing them in this new scenario.