One-shot learning of view-invariant object representations in newborn chicks

Undercompression errors as evidence for conceptual primitives

The Meaning First Approach offers a model of the relation between thought and language that includes a Generator and a Compressor. The Generator build non-linguistic thought structures and the Compressor is responsible for its articulation through three processes: structure-preserving linearization, lexification, and compression via non-articulation of concepts when licensed. One goal of this paper is to show that a range of phenomena in child language can be explained in a unified way within the Meaning First Approach by the assumption that children differ from adults with respect to compression and, specifically, that they may undercompress in production, an idea that sets a research agenda for the study of language acquisition. We focus on dependencies involving pronouns or gaps in relative clauses and wh-questions, multi-argument verbal concepts, and antonymic concepts involving negation or other opposites. We present extant evidence from the literature that children produce undercompression errors (a type of commission errors) that are predicted by the Meaning First Approach. We also summarize data that children's comprehension ability provides evidence for the Meaning First Approach prediction that decompression should be challenging, when there is no 1-to-1 correspondence.

The best game in town: The reemergence of the language-of-thought hypothesis across the cognitive sciences

Mental representations remain the central posits of psychology after many decades of scrutiny. However, there is no consensus about the representational format(s) of biological cognition. This paper provides a survey of evidence from computational cognitive psychology, perceptual psychology, developmental psychology, comparative psychology, and social psychology, and concludes that one type of format that routinely crops up is the language-of-thought (LoT). We outline six core properties of LoTs: (i) discrete constituents; (ii) role-filler independence; (iii) predicate-argument structure; (iv) logical operators; (v) inferential promiscuity; and (vi) abstract content. These properties cluster together throughout cognitive science. Bayesian computational modeling, compositional features of object perception, complex infant and animal reasoning, and automatic, intuitive cognition in adults all implicate LoT-like structures. Instead of regarding LoT as a relic of the previous century, researchers in cognitive science and philosophy-of-mind must take seriously the explanatory breadth of LoT-based architectures. We grant that the mind may harbor many formats and architectures, including iconic and associative structures as well as deep-neural-network-like architectures. However, as computational/representational approaches to the mind continue to advance, classical compositional symbolic structures - that is, LoTs - only prove more flexible and well-supported over time.

Visual Perception of Photographs of Rotated 3D Objects in Goldfish (Carassius auratus)

This study examined goldfishes’ ability to recognize photographs of rotated 3D objects. Six goldfish were presented with color photographs of a plastic model turtle and frog at 0° in a two-alternative forced-choice task. Fish were tested with stimuli at 0°, 90°, 180°, and 270° rotated in the picture plane and two depth planes. All six fish performed significantly above chance at all orientations in the three rotation planes tested. There was no significant difference in performance as a function of aspect angle, which supported viewpoint independence. However, fish were significantly faster at 180° than at +/−90°, so there is also evidence for viewpoint-dependent representations. These fish subjects performed worse overall in the current study with 2D color photographs (M = 88.0%) than they did in our previous study with 3D versions of the same turtle and frog stimuli (M = 92.6%), although they performed significantly better than goldfish in our two past studies presented with black and white 2D stimuli (M = 67.6% and 69.0%). The fish may have relied on color as a salient cue. This study was a first attempt at examining picture-object recognition in fish. More work is needed to determine the conditions under which fish succeed at object constancy tasks, as well as whether they are capable of perceiving photographs as representations of real-world objects.

Ten Questions for a Theory of Vision

By and large, the remarkable progress in visual object recognition in the last few years has been fueled by the availability of huge amounts of labelled data paired with powerful, bespoke computational resources. This has opened the doors to the massive use of deep learning, which has led to remarkable improvements on new challenging benchmarks. While acknowledging this point of view, in this paper I claim that the time has come to begin working towards a deeper understanding of visual computational processes that, instead of being regarded as applications of general purpose machine learning algorithms, are likely to require tailored learning schemes. A major claim of in this paper is that current approaches to object recognition lead to facing a problem that is significantly more difficult than the one offered by nature. This is because of learning algorithms that work on images in isolation, while neglecting the crucial role of temporal coherence. Starting from this remark, this paper raises ten questions concerning visual computational processes that might contribute to better solutions to a number of challenging computer vision tasks. While this paper is far from being able to provide answers to those questions, it contains some insights that might stimulate an in-depth re-thinking in object perception, while suggesting research directions in the control of object-directed action.