Avatar

Andrew Perfors

Professor

University of Melbourne

I am Director of the Complex Human Data Hub at the University of Melbourne School of Psychological Sciences, where I run the Computational Cognitive Science Lab along with Charles Kemp. My research focuses on quantitative approaches to higher-order cognition: concepts, language, decision-making, information and misinformation transmission, and cultural and social evolution and change.

I use a combination of computational models and experiments to understand the why and the what within these topics. What goals are human learners and reasoners trying to achieve in particular situations? What constraints (cognitive, informational, environmental, social) do they operate under? How do these factors shape their behaviour, both on an individual and a group level? How does this behaviour change the informational environment and what there is to learn?

Interests

  • Categorisation, decision-making, learning, language
  • Evolution of social and cultural systems
  • Information and misinformation transmission

Education

  • PhD in Brain & Cognitive Sciences, 2008

    Massachusetts Institute of Technology

  • MA in Linguistics, 2000

    Stanford University

  • BS in Symbolic Systems, 1999

    Stanford University

Research

Concepts

Concept knowledge, inference, and learning

Culture

Social learning and its effects on information transmission and culture

Decision-making

Making choices in a complex world

Language

Language acquisition, representation, and evolution

Miscellaneous

Other projects, mostly focused on meta-science questions

Featured Publications

Ambiguity attitudes in qualitative contexts: The role of prior beliefs

People are averse to ambiguity in a lot of real-world reasoning situations, in part because they are pessimistic about the outcomes
PDF Project DOI PsyArXiv OSF Published

Base rate neglect and conservatism in probabilistic reasoning: Insights from eliciting full distributions

We disentangled base rate neglect and conservatism and found that most people show both but there is a lot of heterogeneity
PDF Project DOI PsyArXiv Supplemental GitHub Published OSF

What do our sampling assumptions affect: How we encode data or how we reason from it?

People need to know how data were generated when they encode it; they can’t revise it later if their assumptions were wrong
PDF Project DOI PsyArXiv Published

The role of stimulus-specific perceptual fluency in statistical learning

Humans have the ability to learn surprisingly complicated statistical information in a variety of modalities and situations, often …
PDF Project DOI Published PsyArXiv Github

Socially situated transmission: The bias to transmit negative information is moderated by the social context

In a transmission chain experiment, we show that negative information is transmitted more easily between people than positive …
PDF Project Published OSF

A large, longitudinal audiovisual dataset recorded from the infant’s perspective

I put a camera on my toddler’s head and this is the result
PDF Dataset Project Project PsyArXiv OSF Published

Social meta-inference and the evidentiary value of consensus

In a twitter-like experimental environment, we show that people are more influenced by the number of distinct posts than the number of …
PDF Project Project Published

Visual and affective multimodal models of word meaning in language and mind

Demonstrates that models that incorporate visual and affective information capture human representations better than models built from …
PDF Dataset Project Project PsyArXiv Published

The evolution of category systems within and between learners

Provides theoretical and experimental evidence for the idea that individuals search more of the problem space than cultural evolution
PDF Project Published

Cross-situational learning in a Zipfian environment

Shows that cross-situational word learning is improved when words and meanings are distributed in a Zipfian manner.
PDF Project OSF PsyArXiv Published

The "Small World of Words" English word association norms for over 12,000 cue words

Presents the largest English-language word association dataset in the world
PDF Dataset Project Project PsyArXiv Published

Even arbitrary norms influence moral decision-making

Shows that people are affected by norms even if they know they are completely arbitrary
PDF Project OSF Published

Do additional features help or hurt category learning? The curse of dimensionality in human learners

Shows that people are only afflicted by the curse of dimensionality for rule-based categories, not family resemblance ones.
PDF Project OSF PsyArXiv Published

Sample size, number of categories and sampling assumptions: Exploring some differences between categorization and generalization

Points out that people make opposing inferences in categorisation and generalisation tasks, and suggests that it is because of …
PDF Project OSF PsyArXiv Published

When extremists win: Cultural transmission via iterated learning when populations are heterogeneous

Shows that in heterogeneous populations, the more extreme reasoners have a disproportionately large effect on the group
PDF Project Project OSF PsyArXiv Published

Bayesian models of cognition revisited: Setting optimality aside and letting data drive psychological theory

Presents a new approach to Bayesian modelling of cognition that removes the interpretation of optimality
PDF Project PsyArXiv Published

Sensitivity to hypothesis size during information search

Shows that, as predicted by a rational analysis, people prefer a positive test strategy more when hypotheses are sparse.
PDF Project PsyArXiv Published

Adult regularization of inconsistent input depends on pragmatic factors

Provides evidence that adults regularise linguistic input in part because of pragmatic and social reasoning
PDF Project Published

Predicting human similarity judgments with distributional models: The value of word associations

Finds that semantic networks built from word association data capture similarity judgments better than state-of-the-art corpus models …
PDF Project Project Published

Leaping to conclusions: Why premise relevance affects argument strength

Demonstrates that premise non-monotonicity can be explained by people’s assumptions about how data are sampled and captured by a …
PDF Project Published PsyArXiv

Language evolution can be shaped by the structure of the world

Shows that iterated learning chains do not necessarily converge to the prior, but can be shaped by the structure of the world too
PDF Project Project PsyArXiv Published

When do memory limitations lead to regularization? An experimental and computational investigation

Uses a combination of modelling and experiments to show that memory limitations at encoding cannot lead to regularisation.
PDF Project Published

Bayesian models of cognition: What's built in after all?

Uses the Bayesian modelling framework to discuss issues of innateness a la Fodor.
PDF Project Published

The learnability of abstract syntactic principles

Uses a Bayesian model of grammar induction to show that a rational learner could infer that language has hierarchical phrase structures …
PDF Project Published

A tutorial introduction to Bayesian models of cognitive development

Gives an intuitive intro to Bayesian models as they can be used to apply to developmental questions.
PDF Project Published

Hypothesis generation, the positive test strategy, and sparse categories

Reveals that positive tests are a rational strategy in a world (like ours) where most hypotheses (and categories) are sparse.
PDF Project PsyArXiv Published

Variability, negative evidence, and the acquisition of verb argument constructions

Presents a model of verb construction acquisition and demonstrates the general principles it uses to resolve the logical problem of …
PDF Project Published

Learning overhypotheses with hierarchical Bayesian models

Demonstrated that Bayesian hierarchical models can capture how humans “learn to learn” new concepts
PDF Project Published

Picking up speed in understanding: Speech processing efficiency and vocabulary growth across the 2nd year

Longitudinal study of infants across the second year showing relationship between word learning and speech processing efficiency
PDF Project Published

Contact

  • University of Melbourne, Parkville, VIC 3010
  • Melbourne Connect level 8