Topic: predictive coding

Develop computational and robotic models of learning, development and social cognition in Jun Tani’s Cognitive Neurorobotics Research Unit. Topics include recurrent networks, predictive coding, active inference, generalisation during development, executive control and working memory. Duties include model and task design, simulation or robotics experiments, analysis and first-author papers. The full-time appointment initially lasts two years; the expected annual salary is JPY 4.6–5.9 million. Recruitment remains open until all positions are filled, with an early start preferred. Follow the emai

Current machine learning systems consume vastly more energy than biological brains. Neuromorphic systems aim to overcome this difference by mimicking the brain’s information coding via discrete voltage spikes. However, it remains unclear how both artificial and natural networks of spiking neurons can learn energy-efficient information processing strategies. Here we propose Predictive Coding Light (PCL), a recurrent hierarchical spiking neural network for unsupervised representation learning. In contrast to previous predictive coding approaches, PCL does not transmit prediction errors to higher

In a first part of the talk, I will present Predictive Coding Light (PCL), a novel unsupervised learning architecture for spiking neural networks. In contrast to conventional predictive coding approaches, which only transmit prediction errors to higher processing stages, PCL learns inhibitory lateral and top-down connectivity to suppress the most predictable spikes and passes a compressed representation of the input to higher processing stages. We show that PCL reproduces a range of biological findings and exhibits a favorable tradeoff between energy consumption and downstream classification p

VIP inhibitory neurons in the cortex receive neuromodulatory and top-down feedback inputs, hence they are well-suited to process behavioral and predictive signals. A recent study found that VIP neurons respond to unexpected image omissions in an image sequence (Garrett et al., 2020). This omission-evoked activity may represent changes in the animal’s behavioral state (e.g., running or arousal), or it may represent a prediction-error signal that could either be stimulus specific, as suggested by the classical predictive coding models (Rao & Ballard, 1992), or stimulus non-specific. We investiga

Head direction (HD) cells track an animal’s head direction in darkness by integrating angular velocity signals, a phenomenon called path integration (PI). Ring attractor models for angular PI have received strong experimental support. To function as integrators, HD circuits require precisely tuned connectivity, which is costly to pass down genetically. This suggests that synaptic plasticity is crucial in setting up these circuits. We propose a network model in which a local, biologically plausible learning rule adjusts synaptic efficacies during development, guided by supervisory allothetic c

Predictive coding represents a promising framework for understanding brain function, postulating that the brain continuously inhibits predictable sensory input, ensuring a preferential processing of surprising elements. A central aspect of this view on cortical computation is its hierarchical connectivity, involving recurrent message passing between excitatory bottom-up signals and inhibitory top-down feedback. Here we use computational modelling to demonstrate that such architectural hard-wiring is not necessary. Rather, predictive coding is shown to emerge as a consequence of energy efficien

Deep learning has made tremendous progress in the last year but it's high computational and memory requirements impose challenges in using deep learning on edge devices. There has been some progress in lowering memory requirements of deep neural networks (for instance, use of half-precision) but there has been minimal effort in developing alternative efficient computational paradigms. Inspired by the brain, Spiking Neural Networks (SNN) provide an energy-efficient alternative to conventional rate-based neural networks. However, SNN architectures that employ the traditional feedforward and feed

At any given moment the brain receives more sensory information than it can use to guide adaptive behaviour, creating the need for mechanisms that promote efficient processing of incoming sensory signals. One way in which the brain might reduce its sensory processing load is to encode successive presentations of the same stimulus in a more efficient form, a process known as neural adaptation. Conversely, when a stimulus violates an expected pattern, it should evoke an enhanced neural response. Such a scheme for sensory encoding has been formalised in predictive coding theories, which propose t

Human vision flaunts a remarkable ability to recognize objects in the surrounding environment even in the absence of complete visual representation of these objects. This process is done almost intuitively and it was not until scientists had to tackle this problem in computer vision that they noticed its complexity. While current advances in artificial vision systems have made great strides exceeding human level in normal vision tasks, it has yet to achieve a similar robustness level. One cause of this robustness is the extensive connectivity that is not limited to a feedforward hierarchical p

Applying the principles of predictive coding to reading has successfully increased our understanding of the neuro-cognitive processes underlying the extraction of meaning from text. Central is integrating prior knowledge with the sensory input to highlight the informative parts of a visual percept, thus increasing the efficiency of the representation and processing in the visual pathway. Current computer vision models (e.g., deep learning models) can easily predict letter identities; however, their inner workings remain opaque, making it extremely hard to infer how the model implements the t

Perception of external stimuli slowly drifts between phases of increased and decreased sensitivity in sustained attention tasks. This phenomenon may represent a deliberate cognitive strategy to balance evidence accumulation against evidence gathering according to the predictive coding hypothesis, or it might simply indicate computational bottlenecks. Here, we claim that dynamical systems theory and large-scale network modeling may offer crucial insights to adjudicate this debate. We attempt to harmonize brain dynamics with dual processing modes by suggesting a link with another aspect of neuro

The ability to robustly recognize strings of letters, a cornerstone of reading, was observed in Baboons and Pigeons despite their lack of phonological and semantic knowledge. Here, we apply a comparative modeling approach to investigate the neuro-cognitive basis of Human, Baboon, and Pigeon orthographic decision behavior, addressing whether phylogenetic relatedness entails similar underlying neuro-cognitive phenotypes. We use the highly transparent Speechless Reader Model (SLR), which assumes letter string recognition based on widely accepted computational principles of predictive coding so th

Our ability to perceive speech remains resilient to variations in voice, rate, and temporal interruptions. While inference models addressed the first two [1], the computational principles by which speech understanding remains impervious to temporal restructuring are largely unexplored. Previous studies indicated intriguing recoveries in comprehension when speech is temporally modulated [2-4]. We show that predictive coding, constrained by endogenous rhythms, accounts for these quizzical results and enables robust lexical speech recovery. We build upon the major hypotheses that the rhythmic st