Generalization, similarity, and Bayesian inference.

By Josh Tenenbaum, T. Griffiths. Behavioral and Brain Sciences, 2001

788 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Cognitive Neuroscience

View Paper: https://www.semanticscholar.org/paper/caa359864efb289d43afc50f2020cf38dcf6fad2

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Cognitive Neuroscience

Citations in the corpus

When is a linguistic metaphor a conceptual metaphor (2009, background): “Tenenbaum and Gri'ths (2001) proposed a Bayesian model according to which the similarity of two items is computed in terms of the probability that they are members of the same category (i.e. drawn from the same statistical distribution).”
What Does It Mean to Span Cultural Boundaries? Variety and Atypicality in Cultural Consumption (2016, background): “Following the foundational work of Shepard (1987) (see also Chater and Vitányi 2003; Tenenbaum and Griffiths 2002), we posit a negative exponential relationship between perceived sociocultural distance and similarity: sim i j d i j, exp , , .”
Learning Causal Laws (2003, neutral): “If we make the simplifying assumption that each network consistent with a grammar Gi is equally likely to be generated by it, the likelihoods P (N |Gi) follow the size principle (Tenenbaum & Griffiths, 2001): P (N |Gi) = 1/size(Gi), for any network N consistent with Gi.”
Word learning as Bayesian inference. (2007, methodology): “…is intended only as a simple but tractable first approximation to the hypothesis space people could adopt for learning common object labels; it is not intended to be the only source of object-label hypotheses or to represent the structure of hypothesis spaces for learning other kinds of words.”
A Century of Gestalt Psychology in Visual Perception II. Conceptual and Theoretical Foundations (2012, background): “…(sometimes referred to as Bayes’ Occam), essentially because larger families of hypotheses (involving more parameters, and in this sense inherently more complex) must assign a lower prior probability to each individual hypothesis (Jeffreys, 1961; D. J. C. MacKay, 2003; Tenenbaum & Griffiths, 2001).”
Object perception as Bayesian inference. (2004, methodology): “Tenenbaum (1999, 2001) provides a Bayesian synthesis of two theories of generalization (similarity-like and rule-like), and provides a computational framework that helps to explain rapid category generalization.”
Three Exemplars Allow at Least Some Linguistic Generalizations: Implications for Generalization Mechanisms and Constraints (2008, positive): “What sort of mechanism requires a small handful of examples, but will not generalize with a single example? To our knowledge, the only currently proposed generalization mechanisms for which the number of input examples is key are Bayesian hypothesis selection mechanisms (e.g., Tenenbaum & Griffiths”
Structure learning in action (2010, neutral): “Evidence for structural learning in motor control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .”
Learning Semantic Image Representations at a Large Scale (2014, background): “Given a hypothesis, the conditional probability P (yi|h) follows from assuming strong sampling [105] and the size principle, thus the conditional probability is defined as follows:”