Topics in Cognitive Science最新文献

Some of the required characteristics for a true machine theory of mind (MToM) include the ability to (1) reproduce the full diversity of human thought and behavior, (2) develop a personalized model of an individual with very limited data, and (3) provide an explanation for behavioral predictions grounded in the cognitive processes of the individual. We propose that a certain class of cognitive models provide an approach that is well suited to meeting those requirements. Being grounded in a mechanistic framework like a cognitive architecture such as ACT-R naturally fulfills the third requirement by mapping behavior to cognitive mechanisms. Exploiting a modeling paradigm such as instance-based learning accounts for the first requirement by reflecting variations in individual experience into a diversity of behavior. Mechanisms such as knowledge tracing and model tracing allow a specific run of the cognitive model to be aligned with a given individual behavior trace, fulfilling the second requirement. We illustrate these principles with a cognitive model of decision-making in a search and rescue task in the Minecraft simulation environment. We demonstrate that cognitive models personalized to individual human players can provide the MToM capability to optimize artificial intelligence agents by diagnosing the underlying causes of observed human behavior, projecting the future effects of potential interventions, and managing the adaptive process of shaping human behavior. Examples of the inputs provided by such analytic cognitive agents include predictions of cognitive load, probability of error, estimates of player self-efficacy, and trust calibration. Finally, we discuss implications for future research and applications to collective human-machine intelligence.

With the rise of artificial intelligence (AI) and the desire to ensure that such machines work well with humans, it is essential for AI systems to actively model their human teammates, a capability referred to as Machine Theory of Mind (MToM). In this paper, we introduce the inner loop of human-machine teaming expressed as communication with MToM capability. We present three different approaches to MToM: (1) constructing models of human inference with well-validated psychological theories and empirical measurements; (2) modeling human as a copy of the AI; and (3) incorporating well-documented domain knowledge about human behavior into the above two approaches. We offer a formal language for machine communication and MToM, where each term has a clear mechanistic interpretation. We exemplify the overarching formalism and the specific approaches in two concrete example scenarios. Related work that demonstrates these approaches is highlighted along the way. The formalism, examples, and empirical support provide a holistic picture of the inner loop of human-machine teaming as a foundational building block of collective human-machine intelligence.

This paper explores a framework for defining artificial intelligence (AI) that adapts to individuals within a group, and discusses the technical challenges for collaborative AI systems that must work with different human partners. Collaborative AI is not one-size-fits-all, and thus AI systems must tune their output based on each human partner's needs and abilities. For example, when communicating with a partner, an AI should consider how prepared their partner is to receive and correctly interpret the information they are receiving. Forgoing such individual considerations may adversely impact the partner's mental state and proficiency. On the other hand, successfully adapting to each person's (or team member's) behavior and abilities can yield performance benefits for the human-AI team. Under this framework, an AI teammate adapts to human partners by first learning components of the human's decision-making process and then updating its own behaviors to positively influence the ongoing collaboration. This paper explains the role of this AI adaptation formalism in dyadic human-AI interactions and examines its application through a case study in a simulated navigation domain.

Artificial Intelligence (AI) powered machines are increasingly mediating our work and many of our managerial, economic, and cultural interactions. While technology enhances individual capability in many ways, how do we know that the sociotechnical system as a whole, consisting of a complex web of hundreds of human-machine interactions, is exhibiting collective intelligence? Research on human-machine interactions has been conducted within different disciplinary silos, resulting in social science models that underestimate technology and vice versa. Bringing together these different perspectives and methods at this juncture is critical. To truly advance our understanding of this important and quickly evolving area, we need vehicles to help research connect across disciplinary boundaries. This paper advocates for establishing an interdisciplinary research domain-Collective Human-Machine Intelligence (COHUMAIN). It outlines a research agenda for a holistic approach to designing and developing the dynamics of sociotechnical systems. In illustrating the kind of approach, we envision in this domain, we describe recent work on a sociocognitive architecture, the transactive systems model of collective intelligence, that articulates the critical processes underlying the emergence and maintenance of collective intelligence and extend it to human-AI systems. We connect this with synergistic work on a compatible cognitive architecture, instance-based learning theory and apply it to the design of AI agents that collaborate with humans. We present this work as a call to researchers working on related questions to not only engage with our proposal but also develop their own sociocognitive architectures and unlock the real potential of human-machine intelligence.

Phenomenal experiences of immaterial spiritual beings-hearing the voice of God, seeing the spirit of an ancestor-are a valuable and largely untapped resource for the field of cognitive science. Such experiences, we argue, are experiences of the mind, tied to mental models and cognitive-epistemic attitudes about the mind, and thus provide a striking example of how, with the right combination of mental models and cognitive-epistemic attitudes, one's own thoughts and inner sensations can be experienced as coming from somewhere or someone else. In this paper, we present results from a large-scale study of U.S. adults (N = 1779) that provides new support for our theory that spiritual experiences are facilitated by a dynamic interaction between mental models and cognitive-epistemic attitudes: A person is more likely to hear God speak if they have the epistemic flexibility and cultural support to shift, temporarily, away from a mundane model of mind into a more "porous" way of thinking and being. This, in turn, lays the foundation for a meditation on how mental models and cognitive-epistemic attitudes might also interact to facilitate other phenomena of interest to cognitive science, such as fiction writing and scientific discovery.

In the modern world, many important tasks have become too complex for a single unaided individual to manage. Teams conduct some safety-critical tasks to improve task performance and minimize the risk of error. These teams have traditionally consisted of human operators, yet, nowadays, artificial intelligence and machine systems are incorporated into team environments to improve performance and capacity. We used a computerized task modeled after a classic arcade game to investigate the performance of human-machine and human-human teams. We manipulated the group conditions between team members; sometimes, they were instructed to collaborate, compete, or work separately. We evaluated players' performance in the main task (gameplay) and, in post hoc analyses, participant behavioral patterns to inform group strategies. We compared game performance between team types (human-human vs. human-machine) and group conditions (competitive, collaborative, independent). Adapting workload capacity analysis to human-machine teams, we found performance under both team types and all group conditions suffered a performance efficiency cost. However, we observed a reduced cost in collaborative over competitive teams within human-human pairings, but this effect was diminished when playing with a machine partner. The implications of workload capacity analysis as a powerful tool for human-machine team performance measurement are discussed.

Socio-cognitive theory conceptualizes individual contributors as both enactors of cognitive processes and targets of a social context's determinative influences. The present research investigates how contributors' metacognition or self-beliefs, combine with others' views of themselves to inform collective team states related to learning about other agents (i.e., transactive memory systems) and forming social attachments with other agents (i.e., collective team identification), both important teamwork states that have implications for team collective intelligence. We test the predictions in a longitudinal study with 78 teams. Additionally, we provide interview data from industry experts in human-artificial intelligence teams. Our findings contribute to an emerging socio-cognitive architecture for COllective HUman-MAchine INtelligence (i.e., COHUMAIN) by articulating its underpinnings in individual and collective cognition and metacognition. Our resulting model has implications for the critical inputs necessary to design and enable a higher level of integration of human and machine teammates.

The use of terms like "collaboration" and "co-workers" to describe interactions between human beings and certain artificial intelligence (AI) systems has gained significant traction in recent years. Yet, it remains an open question whether such anthropomorphic metaphors provide either a fertile or even a purely innocuous lens through which to conceptualize designed commercial products. Rather, a respect for human dignity and the principle of transparency may require us to draw a sharp distinction between real and faux peers. At the heart of the concept of collaboration lies the assumption that the collaborating parties are (or behave as if they are) of similar status: two agents capable of comparable forms of intentional action, moral agency, or moral responsibility. In application to current AI systems, this not only seems to fail ontologically but also from a socio-political perspective. AI in the workplace is primarily an extension of capital, not of labor, and the AI "co-workers" of most individuals will likely be owned and operated by their employer. In this paper, we critically assess both the accuracy and desirability of using the term "collaboration" to describe interactions between humans and AI systems. We begin by proposing an alternative ontology of human-machine interaction, one which features not two equivalently autonomous agents, but rather one machine that exists in a relationship of heteronomy to one or more human agents. In this sense, while the machine may have a significant degree of independence concerning the means by which it achieves its ends, the ends themselves are always chosen by at least one human agent, whose interests may differ from those of the individuals interacting with the machine. We finally consider the motivations and risks inherent to the continued use of the term "collaboration," exploring its strained relation to the concept of transparency, and consequences for the future of work.

Artificial social intelligence (ASI) agents have great potential to aid the success of individuals, human-human teams, and human-artificial intelligence teams. To develop helpful ASI agents, we created an urban search and rescue task environment in Minecraft to evaluate ASI agents' ability to infer participants' knowledge training conditions and predict participants' next victim type to be rescued. We evaluated ASI agents' capabilities in three ways: (a) comparison to ground truth-the actual knowledge training condition and participant actions; (b) comparison among different ASI agents; and (c) comparison to a human observer criterion, whose accuracy served as a reference point. The human observers and the ASI agents used video data and timestamped event messages from the testbed, respectively, to make inferences about the same participants and topic (knowledge training condition) and the same instances of participant actions (rescue of victims). Overall, ASI agents performed better than human observers in inferring knowledge training conditions and predicting actions. Refining the human criterion can guide the design and evaluation of ASI agents for complex task environments and team composition.