Social biases are prevalent in everyday social interactions, but they are often expressed in subt... more Social biases are prevalent in everyday social interactions, but they are often expressed in subtle ways that can make them difficult to detect. Yet, intuitively, people can often recognize when they are the subject of a bias, even in the absence of any overt behavior. How do we do this? While much research has focused on the negative consequences of being the subject of a bias, less is known about the cognitive mechanisms that allow people to explicitly detect biases in the first place. In this paper, we propose an account of bias detection which is grounded on mental state representations. We propose that people infer biases by detecting a gap between expected unbiased
Human social interactions require understanding and predicting other people's behavior. A growing... more Human social interactions require understanding and predicting other people's behavior. A growing body of work has found that these inferences are structured around an assumption that agents act rationally and efficiently in space. While powerful, this view treats action understanding in a vacuum, ignoring that much social inference happens in the context of familiar, hierarchically structured events (e.g.: buying groceries, ordering in a restaurant). We propose that social and world knowledge is critical for efficiently interpreting behavior and test this idea through a simple blockbuilding paradigm, where participants infer an agent's sub-task (study 1a), next action (study 1b), and higher-level goal (study 1c), from very sparse observations. We compare these inferences against a Bayesian model of goal inference that exploits task structure to interpret agents' actions. This model fit participant judgments with high quantitative accuracy, highlighting how world knowledge may help support social inferences in a rich and powerful way.
We investigate how humans infer the rich internal structure of social collectives from patterns o... more We investigate how humans infer the rich internal structure of social collectives from patterns of interactions between agents. We propose a computational model of this process which integrates a domain-general statistical learning mechanism with, domain-specific knowledge about social contexts (i.e.: "intuitive sociologies"). We test our model in two experiments where participants observe a sequence of animated interactions between agents, and then assign the agents to groups according to their role or type within the social collective. Crucially, the two experiments depict different types of social interactions which reflect different types of underlying social structures. The patterns of correspondence between model predictions and human data support our account, and demonstrate the importance of both general statistical learning and specific social knowledge when reasoning about social collectives.
How do we know what babies know? The limits of inferring cognitive representations from visual fixation data
Philosophical Psychology, 2020
Most infant cognitive studies use visual fixation time as the measure of interest. There are, how... more Most infant cognitive studies use visual fixation time as the measure of interest. There are, however, some serious methodological and theoretical concerns regarding what these studies reveal about infant cognition and how their results ought to be interpreted. We propose a Bayesian modeling framework which helps address these concerns. This framework allows us to more precisely formulate hypotheses about infants’ cognitive representations, formalize “linking hypotheses” that relate infants’ visual fixation behavior with stimulus complexity, and better determine what questions a given experiment can and cannot answer.
In this paper, we propose a flexible modeling framework for studying the role of perception in la... more In this paper, we propose a flexible modeling framework for studying the role of perception in language learning and language evolution. This is achieved by augmenting some novel and some existing evolutionary signaling game models with existing techniques in machine learning and cognitive science. The result is a “grounded” signaling game in which agents must extract relevant information from their environment via a cognitive processing mechanism, then learn to communicate that information with each other. The choice of cognitive processing mechanism is left as a free parameter, allowing the model to be tailored to a wide variety of problems and tasks. We present results from simulations using both a Bayesian perception model and a neural network based perception model, which demonstrate how perception can “preprocess” environmental data in a way that is well suited for communication. Lastly, we discuss how the model can be extended to study other roles that perception may play in ...
In this paper we give an introduction to nonstandard analysis, starting with an ultrapower constr... more In this paper we give an introduction to nonstandard analysis, starting with an ultrapower construction of the hyperreals. We then demonstrate how theorems in standard analysis “transfer over” to nonstandard analysis, and how theorems in standard analysis can be proven using theorems in nonstandard analysis.
By associating a subfield of R to a set of points P0 ⊆ R2, geometric properties of ruler and comp... more By associating a subfield of R to a set of points P0 ⊆ R2, geometric properties of ruler and compass constructions on P0 can be understood algebraically, creating a powerful tool for proving the possibility or impossibility of certain constructions. In this paper, field theory will be used to prove the impossibility of doubling the cube and squaring the circle, and will be used in studying the constructibility of regular n-gons.
I can tell you know a lot, although I'm not sure what: Modeling broad epistemic inference from minimal action
Inferences about other people's knowledge and beliefs are central to social interaction. In m... more Inferences about other people's knowledge and beliefs are central to social interaction. In many situations, however, it's not possible to be sure what other people know because their behavior is consistent with a range of potential epistemic states. Nonetheless, this behavior can give us coarse intuitions about how much someone might know, even if we cannot pinpoint the exact nature of this knowledge. We present a computational model of this kind of broad epistemic-state inference, centered on the expectation that agents maximize epistemic utilities. We evaluate our model in a graded inference task where people had to infer how much an agent knew based on the actions they chose. Critically, the agent's behavior was always under-determined, but nonetheless contained information about how much knowledge they possessed. Our model captures nuanced patterns in participant judgments, revealing a quantitative capacity to infer amorphous knowledge from minimal behavioral evidence.
Reasoning about social preferences with uncertain beliefs
We propose a computational model of social preference judgments that accounts for the degree of a... more We propose a computational model of social preference judgments that accounts for the degree of an agents’ uncertainty about the preferences of others. Underlying this model is the principle that, in the face of social uncertainty, people interpret social agents’ behavior under an assumption of expected utility maximization. We evaluate our model in two experiments which each test a different kind of social preference reasoning: predicting social choices given information about social preferences, and inferring social preferences after observing social choices. The results support our model and highlight how un- certainty influences our social judgments.
We propose a framework for pragmatic reliability-in-the-limit criteria, extending the epistemic r... more We propose a framework for pragmatic reliability-in-the-limit criteria, extending the epistemic reliability framework (Kelly 1996). We identify some common scientific contexts which complicate the application or interpretation of epistemic reliability criteria, drawing heavily from economics for illustrative examples. We then propose an extension of the standard framework, where inquiry is constrained by both epistemic and non-epistemic factors. This provides analogous notions of pragmatic underdetermination and pragmatic reliability with respect to a particular goal, as well as a principled method for extracting solvable problems from unsolvable ones.
Almost Fair: Conjoint Measurement Theory and Score-Based Bargaining Solutions
Communications in Computer and Information Science, 2016
A bargaining problem is a cooperative game in which players are permitted to negotiate before the... more A bargaining problem is a cooperative game in which players are permitted to negotiate before the game is played. Bargaining theory can be used to economic interactions such as union negotiations, international trade agreements, and duopolies. A general theory of bargaining games thus has a wide application to many areas of economics and political science.
Uploads
Papers by Isaac Davis