A little labeling goes a long way: Semi‐supervised learning in infancy

Abstract

There is considerable evidence that labeling supports infants' object categorization. Yet in daily life, most of the category exemplars that infants encounter will remain unlabeled. Inspired by recent evidence from machine learning, we propose that infants successfully exploit this sparsely labeled input through "semi-supervised learning." Providing only a few labeled exemplars leads infants to initiate the process of categorization, after which they can integrate all subsequent exemplars, labeled or unlabeled, into their evolving category representations. Using a classic novelty preference task, we introduced 2-year-old infants (n = 96) to a novel object category, varying whether and when its exemplars were labeled. Infants were equally successful whether all exemplars were labeled (fully supervised condition) or only the first two exemplars were labeled (semi-supervised condition), but they failed when no exemplars were labeled (unsupervised condition). Furthermore, the timing of the labeling mattered: when the labeled exemplars were provided at the end, rather than the beginning, of familiarization (reversed semi-supervised condition), infants failed to learn the category. This provides the first evidence of semi-supervised learning in infancy, revealing that infants excel at learning from exactly the kind of input that they typically receive in acquiring real-world categories and their names.

Do Preverbal Infants Understand Discrete Facial Expressions of Emotion?

“…However, infants only formed a superordinate category (vehicles) when a noun was presented. Similar facilitative labeling effects have subsequently been found with basic-level categories (e.g., LaTourrette & Waxman, 2019), novel objects (e.g., Fulkerson & Haaf, 2006), and other object properties (e.g., spatial relationships; Casasola, Bhagwat, & Burke, 2009).…”

Defining the Role Of Language in Infants' Object Categorization with Eye-tracking Paradigms

“…Using the protocol above, we ran two experiments 22 . Analyses were conducted with the eyetrackingR package 23 , and the data and code are available at https://github.com/sandylat/ ssl-in-infancy.…”

Naming guides how 12-month-old infants encode and remember objects

“…Comparison is instrumental to identifying both similarities (e.g., for infants in the Consistent Name condition) and differences (e.g., for infants in the Distinct Names condition) (27,28). Moreover, with two consistently named exemplars, infants successfully begin to identify objects' commonalities (4,29,30). However, by Training Trials 3 and 4, infants' encoding of the training objects should vary as function of condition.…”