Analyzing Machine-Learned Representations: A Natural Language Case Study

As modern deep networks become more complex, and get closer to human-like capabilities in certain domains, the question arises as to how the representations and decision rules they learn compare to the ones in humans. In this work, we study representations of sentences in one such artificial system for natural language processing. We first present a diagnostic test dataset to examine the degree of abstract composable structure represented. Analyzing performance on these diagnostic tests indicates a lack of systematicity in representations and decision rules, and reveals a set of heuristic strategies. We then investigate the effect of training distribution on learning these heuristic strategies, and we study changes in these representations with various augmentations to the training set. Our results reveal parallels to the analogous representations in people. We find that these systems can learn abstract rules and generalize them to new contexts under certain circumstances—similar to human zero-shot reasoning. However, we also note some shortcomings in this generalization behavior—similar to human judgment errors like belief bias. Studying these parallels suggests new ways to understand psychological phenomena in humans as well as informs best strategies for building artificial intelligence with human-like language understanding.     

Applying the Schlossberg 4S Transition Model to Retired University Faculty: Does It Fit?

The authors surveyed university faculty to investigate whether the Schlossberg 4S Transition Model was a useful way to look at retirement adjustment. Participants were asked questions regarding aspects of situation, self, support, and strategies before and after retirement. Results are described in relation to overall retirement satisfaction.     

The Great Divorce: Ethics and Identity

Significant epistemological shifts took place during the modern period that led to the exclusion of ethical constructs of the human self. Subsequently, ethics or morality was systematically divorced from reason and human identity. This has left modern psychologies with impoverished languages for understanding morality and ethics, especially as it relates to human identity. Some of these changes and their subsequent effects are chronicled. Toward this end, three unique influences of the modern era are surveyed. First, it is argued that the nature of reasoning and rationality underwent a significant shift in definition and emphasis, becoming autonomous and secular. Second, it is suggested that the scope of inquiry was restricted to the natural order, creating an immanentization of knowledge. Lastly, it is argued that modernity ushered in a promotion of and preoccupation with the individual subject. The context of each influence are explored, along with a subsection in each describing how the effects of these impacted the growing divorce between ethics and human identity.     

Psychology in Pursuit of Justice: The Lives and Works of Emmanuel Levinas and Ignacio Martín-Baró

Psychological research and theory are inherently political and social practices. As such, these authors assert that social justice must be a primary concern of the psychological sciences. In case study format, the authors use the exceptional lives and works of continental philosopher Emmanuel Levinas and social psychologist Ignacio Martín-Baró to exemplify the recognition that justice must be centrally infused into psychological thought and practice. The life and work of philosopher Levinas is characterized by a concern for how Western philosophy has funded constructs of selfhood and identity that have created moral anemia and hampered our responsibility to others. Martín-Baró, a prominent Latin American researcher and educator, found research to be a powerful tool to expose oppressive social systems and widespread deception in an effort to end unjust practices, thereby affecting social change. Biographical material and details concerning their systems of research and thought will be used to illustrate what it means to engage in morally-just research and theorizing within the field of psychology.     

Understanding and Using Informants’ Reporting Discrepancies of Youth Victimization: A Conceptual Model and Recommendations for Research

Discrepancies often occur among informants’ reports of various domains of child and family functioning and are particularly common between parent and child reports of youth violence exposure. However, recent work suggests that discrepancies between parent and child reports predict subsequent poorer child outcomes. We propose a preliminary conceptual model (Discrepancies in Victimization Implicate Developmental Effects [DiVIDE]) that considers how and why discrepancies between parents’ and youths’ ratings of child victimization may be related to poor adjustment outcomes. The model addresses how dyadic processes, such as the parent–youth relationship and youths’ information management, might contribute to discrepancies. We also consider coping processes that explain why discrepancies may predict increases in youth maladjustment. Based on this preliminary conceptual framework, we offer suggestions and future directions for researchers who encounter conflicting reports of community violence exposure and discuss why the proposed model is relevant to interventions for victimized youths.