How do we map the rapid input of spoken language onto phonological and lexical representations over time? Attempts at psychologically-tractable computational models of spoken word recognition tend either to ignore time or to transform the temporal input into a spatial representation. TRACE, a connectionist model with broad and deep coverage of speech perception and spoken word recognition phenomena, takes the latter approach, using exclusively time-specific units at every level of representation. TRACE reduplicates featural, phonemic, and lexical inputs at every time step in a large memory trace, with rich interconnections (excitatory forward and backward connections between levels and inhibitory links within levels). As the length of the memory trace is increased, or as the phoneme and lexical inventory of the model is increased to a realistic size, this reduplication of time- (temporal position) specific units leads to a dramatic proliferation of units and connections, begging the question of whether a more efficient approach is possible. Our starting point is the observation that models of visual object recognition—including visual word recognition—have grappled with the problem of spatial invariance, and arrived at solutions other than a fully-reduplicative strategy like that of TRACE. This inspires a new model of spoken word recognition that combines time-specific phoneme representations similar to those in TRACE with higher-level representations based on string kernels: temporally independent (time invariant) diphone and lexical units. This reduces the number of necessary units and connections by several orders of magnitude relative to TRACE. Critically, we compare the new model to TRACE on a set of key phenomena, demonstrating that the new model inherits much of the behavior of TRACE and that the drastic computational savings do not come at the cost of explanatory power.

Introduced in the late 1980s for generalization purposes, pruning has now become a staple for compressing deep neural networks. Despite many innovations in recent decades, pruning approaches still face core issues that hinder their performance or scalability. Drawing inspiration from early work in the field, and especially the use of weight decay to achieve sparsity, we introduce Selective Weight Decay (SWD), which carries out efficient, continuous pruning throughout training. Our approach, theoretically grounded on Lagrangian smoothing, is versatile and can be applied to multiple tasks, networks, and pruning structures. We show that SWD compares favorably to state-of-the-art approaches, in terms of performance-to-parameters ratio, on the CIFAR-10, Cora, and ImageNet ILSVRC2012 datasets.

What is the origin of our ability to learn orthographic knowledge? We use deep convolutional networks to emulate the primate's ventral visual stream and explore the recent finding that baboons can be trained to discriminate English words from nonwords [1]. The networks were exposed to the exact same sequence of stimuli and reinforcement signals as the baboons in the experiment, and learned to map real visual inputs (pixels) of letter strings onto binary word/nonword responses. We show that the networks' highest levels of representations were indeed sensitive to letter combinations as postulated in our previous research. The model also captured the key empirical findings, such as generalization to novel words, along with some intriguing inter-individual differences. The present work shows the merits of deep learning networks that can simulate the whole processing chain all the way from the visual input to the response while allowing researchers to analyze the complex representations that emerge during the learning process.

Turning Turing's logic on its head, we used widespread letter-based Turing Tests found on the internet (CAPTCHAs) to shed light on human cognition. We examined the basis of the human ability to solve CAPTCHAs, where machines fail. We asked whether this is due to our use of slow-acting inferential processes that would not be available to machines, or whether fast-acting automatic orthographic processing in humans has superior robustness to shape variations. A masked priming lexical decision experiment revealed efficient processing of CAPTCHA words in conditions that rule out the use of slow inferential processing. This shows that the human superiority in solving CAPTCHAs builds on a high degree of invariance to location and continuous transforms, which is achieved during the very early stages of visual word recognition in skilled readers.

Number sense, a spontaneous ability to process approximate numbers, has been documented in human adults, infants and newborns, and many other animals. Species as distant as monkeys and crows exhibit very similar neurons tuned to specific numerosities. How number sense can emerge in the absence of learning or fine tuning is currently unknown. We introduce a random-matrix theory of self-organized neural states where numbers are coded by vectors of activation across multiple units, and where the vector codes for successive integers are obtained through multiplication by a fixed but random matrix. This cortical implementation of the ‘von Mises' algorithm explains many otherwise disconnected observations ranging from neural tuning curves in monkeys to looking times in neonates and cortical numerotopy in adults. The theory clarifies the origin of Weber–Fechner's Law and yields a novel and empirically validated prediction of multi-peak number neurons. Random matrices constitute a novel mechanism for the emergence of brain states coding for quantity. This article is part of a discussion meeting issue ‘The origins of numerical abilities’.

In the context of automated highway systems (AHS), this work proposes an approach that enables a vehicle to autonomously join a platoon with optimized trajectory in the presence of dynamical traffic obstacles. A notable aspect is the use of Model Predictive Control (MPC) optimization of the planned path, in conjunction with a variant of the Rapidly-exploring Random Trees (RRT*) algorithm for the purpose of platoon formation. This combination efficiently explores the space of possible trajectories, returning trajectories that are smoothened out with respect to the dynamic constraints of the vehicle, while at the same time allowing for real-time implementation. The implementation we propose takes into consideration both localization and mapping through relevant sensors and V2V communication. The complete algorithm is tested over various nominal and worst-case scenarios, qualifying the merits of the proposed methodology.