Spatial relation categorization in infants and deep neural networks

Spatial relations, such as above, below, between, and containment, are important mediators in children's understanding of the world (Piaget, 1954). The development of these relational categories in infancy has been extensively studied (Quinn, 2003) yet little is known about their computational underpinnings. Using developmental tests, we examine the extent to which deep neural networks, pretrained on a standard vision benchmark or egocentric video captured from one baby's perspective, form categorical representations for visual stimuli depicting relations. Notably, the networks did not receive any explicit training on relations. We then analyze whether these networks recover similar patterns to ones identified in development, such as reproducing the relative difficulty of categorizing different spatial relations and different stimulus abstractions. We find that the networks we evaluate tend to recover many of the patterns observed with the simpler relations of "above versus below" or "between versus outside", but struggle to match developmental findings related to "containment". We identify factors in the choice of model architecture, pretraining data, and experimental design that contribute to the extent the networks match developmental patterns, and highlight experimental predictions made by our modeling results. Our results open the door to modeling infants' earliest categorization abilities with modern machine learning tools and demonstrate the utility and productivity of this approach.

Dissociable learning processes, associative theory, and testimonial reviews: A comment on Smith and Church (2018)

Smith and Church (Psychonomic Bulletin & Review, 25, 1565-1584 2018) present a "testimonial" review of dissociable learning processes in comparative and cognitive psychology, by which we mean they include only the portion of the available evidence that is consistent with their conclusions. For example, they conclude that learning the information-integration category-learning task with immediate feedback is implicit, but do not consider the evidence that people readily report explicit strategies in this task, nor that this task can be accommodated by accounts that make no distinction between implicit and explicit processes. They also consider some of the neuroscience relating to information-integration category learning, but do not report those aspects that are more consistent with an explicit than an implicit account. They further conclude that delay conditioning in humans is implicit, but do not report evidence that delay conditioning requires awareness; nor do they present the evidence that conditioned taste aversion, which should be explicit under their account, can be implicit. We agree with Smith and Church that it is helpful to have a clear definition of associative theory, but suggest that their definition may be unnecessarily restrictive. We propose an alternative definition of associative theory and briefly describe an experimental procedure that we think may better distinguish between associative and non-associative processes.

Evidence against interactive effects on articulation in Javanese verb paradigms

In interactive models of speech production, wordforms that are related to a target form are co-activated during lexical planning, and co-activated wordforms can leave phonetic traces on the target. This mechanism has been proposed to account for phonetic similarities among morphologically related wordforms. We test this hypothesis in a Javanese verb paradigm. In Javanese, one class of verbs is inflected by nasalizing an initial voiceless obstruent: one form of each word begins with a nasal, while its otherwise identical relative begins with a voiceless obstruent. We predict that if morphologically related forms are co-activated during production, the nasal-initial forms of these words should show phonetic traces of their obstruent-initial forms, as compared to nasal-initial wordforms that do not alternate. Twenty-seven native Javanese speakers produced matched pairs of alternating and non-alternating wordforms. Based on an acoustic analysis of nasal resonance and closure duration, we present good evidence against the original hypothesis: We find that the alternating nasals are phonetically identical to the non-alternating ones on both measures. We argue that interactive effects during lexical planning do not offer the best account for morphologically conditioned phonetic similarities. We discuss an alternative involving competition between phonotactic constraints and word-specific phonological structures.

NEURAL Networks and Consumer Behavior: NEURAL Models, Logistic Regression, and the Behavioral Perspective Model

This paper investigates the ability of connectionist models to explain consumer behavior, focusing on the feedforward neural network model, and explores the possibility of expanding the theoretical framework of the Behavioral Perspective Model to incorporate connectionist constructs. Numerous neural network models of varying complexity are developed to predict consumer loyalty as a crucial aspect of consumer behavior. Their performance is compared with the more traditional logistic regression model and it is found that neural networks offer consistent advantage over logistic regression in the prediction of consumer loyalty. Independently determined Utilitarian and Informational Reinforcement variables are shown to make a noticeable contribution to the explanation of consumer choice. The potential of connectionist models for predicting and explaining consumer behavior is discussed and routes for future research are suggested to investigate the predictive and explanatory capacity of connectionist models, such as neural network models, and for the integration of these into consumer behavior analysis within the theoretical framework of the Behavioral Perspective Model.

Integrating SLAM with existing evidence: Comment on Walker and Hickok (2015)

Walker and Hickok (Psychonomic Bulletin & Review doi:10.3758/s13423-015-0903-7, 2015) used simulations to compare a novel proposal, the semantic-lexical-auditory-motor model (SLAM), to an existing account of speech production, the two-step interactive account (TSIA; Foygel & Dell, Journal of Memory and Language, 43:182-216, doi:10.1006/jmla.2000.2716, 2000). This commentary critically examines their assessment of SLAM. The cases in which SLAM outperforms TSIA largely reflect SLAM's ability to (poorly) approximate an existing theory of speech production incorporating two stages of phonological processing (the lexical + postlexical account). The fact that SLAM and TSIA can exhibit equivalent fits to the overall response distribution of a set of aphasic patients is unsurprising, since previous work has shown that overall response distributions do not reliably discriminate theoretical alternatives. Finally, SLAM inherits issues associated with TSIA's assumption of strong feedback between levels of representation. This suggests that SLAM does not represent an advance over existing theories of speech production.

Interactive effects of explicit emergent structure: a major challenge for cognitive computational modeling

David Marr's (1982) three-level analysis of computational cognition argues for three distinct levels of cognitive information processing-namely, the computational, representational, and implementational levels. But Marr's levels are-and were meant to be-descriptive, rather than interactive and dynamic. For this reason, we suggest that, had Marr been writing today, he might well have gone even farther in his analysis, including the emergence of structure-in particular, explicit structure at the conceptual level-from lower levels, and the effect of explicit emergent structures on the level (or levels) that gave rise to them. The message is that today's cognitive scientists need not only to understand how emergent structures-in particular, explicit emergent structures at the cognitive level-develop but also to understand how they feed back on the sub-structures from which they emerged.

Episodic intertrial learning of younger and older participants: effects of age of acquisition

There is clear evidence of a deficit in episodic memory for older adults compared to younger adults. Using an intertrial technique previous research has investigated whether this deficit can be attributed to a decline in encoding or consolidation. On standard memory tests, these two aspects of memory function can be measured by examining the items forgotten or acquired across multiple learning trials. The present study assessed whether age deficits in episodic memory were affected by stimulus characteristics, specifically age of acquisition (AoA). A standard intertrial design was implemented whereby participants studied word lists over several study-test trials. The stimulus characteristics of AoA were manipulated using a pure-list technique. Our findings showed that older adults demonstrate an overall recall deficit which appeared to be a consequence of both an encoding deficit and consolidation weakness. Earlier-acquired words were recalled significantly better than later-acquired words and this was apparently due to both enhanced encoding and consolidation of earlier- over later-acquired words. The key finding is that older adults show a recall advantage for earlier- compared to later-acquired words over the entire experiment to a greater degree than younger adults.