Humans tend to anticipate events when they synchronize their actions with sound (such as when they clap to music), which has puzzled scientists for decades. What accounts for this anticipation? We review two theoretic...
详细信息
Humans tend to anticipate events when they synchronize their actions with sound (such as when they clap to music), which has puzzled scientists for decades. What accounts for this anticipation? We review two theoretical mechanisms for synchrony: predictive coding and dynamical systems. Both theories are grounded in neural activation patterns, but there are important distinctions. We contrast their assumptions, their computations, and their musical applications to anticipatory synchronization.
An algorithm for reversible data compression based on predictive coding is presented. From the input data, a sequence of integer-valued residuals is generated by a linear or nonlinear operation. The size of the alphab...
详细信息
An algorithm for reversible data compression based on predictive coding is presented. From the input data, a sequence of integer-valued residuals is generated by a linear or nonlinear operation. The size of the alphabet for the residuals is reduced by performing a modular operation on its symbols. The reconstruction process recovers the original data exactly. The modular operation results in a smaller size codebook and prevents data expansion when the source is not matched to the code. It also reduces the entropy of the residuals, which theoretically should result in a higher degree of data compression, but this seems to be of little practical significance.
Learning temporal patterns among primitive speech sequences and being able to control the motor apparatus for effective production of the learned patterns are imperative for speech acquisition in infants. In this pape...
详细信息
ISBN:
(纸本)9781510833135
Learning temporal patterns among primitive speech sequences and being able to control the motor apparatus for effective production of the learned patterns are imperative for speech acquisition in infants. In this paper, we develop a predictive coding model whose objective is to minimize the sensory (auditory) and proprioceptive prediction errors. Temporal patterns are learned by minimizing the former while control is learned by minimizing the latter. The model is learned using a set of synthetically generated syllables, as in other contemporary models. We show that the proposed model outperforms existing ones in learning vocalization classes. It also computes the control/muscle activation which is useful for determining the degree of easiness of vocalization.
This paper discusses representation learning from electroencephalographic (EEG) signal with deep variational predictive coding *** introduce a hierarchical probabilistic network that minimises prediction error at mult...
详细信息
ISBN:
(纸本)9781728185262
This paper discusses representation learning from electroencephalographic (EEG) signal with deep variational predictive coding *** introduce a hierarchical probabilistic network that minimises prediction error at multiple levels of spatio-temporal abstraction. While the lowest layer predicts brain activity directly, higher layers abstract away from the data and predict sequences of the hidden states in lower layers. The network captures both expected and actual uncertainty by relating predicted state posteriors. Each layer minimises (expected) surprise either with or without sampling new evidence from the layer below. This structure motivates both active learning and active inference as means to learn representations. Active learning refers to model parameter exploration which allows to learn regularities, especially when they are stable between trials. Active inference refers to hidden state exploration, a process that enables dynamic inference of the current context using the learned generative model. We train the model on EEG data recorded during free reading and evaluate adaptive EEG prediction in the context of Fixation Related Potentials (FRPs).
This study is focused on the development of the cortex-like visual object recognition system. We propose a general framework, which consists of three hierarchical levels (modules). These modules functionally correspon...
详细信息
ISBN:
(纸本)9781424496365
This study is focused on the development of the cortex-like visual object recognition system. We propose a general framework, which consists of three hierarchical levels (modules). These modules functionally correspond to the V1, V4 and IT areas. Both bottom-up and top-down connections between the hierarchical levels V4 and IT are employed. The higher the degree of matching between the input and the preferred stimulus, the shorter the response time of the neuron. Therefore information about a single stimulus is distributed in time and is transmitted by the waves of spikes. The reciprocal connections and waves of spikes implement predictive coding: an initial hypothesis is generated on the basis of information delivered by the first wave of spikes and is tested with the information carried by the consecutive waves. The development is considered as extraction and accumulation of features in V4 and objects in IT. Once stored a feature can be disposed, if rarely activated. This causes update of feature repository. Consequently, objects in IT are also updated. This illustrates the growing process and dynamical change of topological structures of V4, IT and connections between these areas.
Ward's signal detection theory-based framework elucidates some aspects of interindividual differences in sensitivity, but, we argue, obscures others. Specifically, it disregards the important challenge of inferrin...
详细信息
Ward's signal detection theory-based framework elucidates some aspects of interindividual differences in sensitivity, but, we argue, obscures others. Specifically, it disregards the important challenge of inferring the meaning of sensory inputs. Within Bayesian predictive coding accounts, the meaning is given by inferences to more deeply hidden causes of sensory inputs and is generally the basis for initiating context-appropriate (e.g., social) behavior. As such, when inference of hierarchical causes is hampered, as accounts of autism based on deficient precision estimation imply, a form of hyporesponsivity can emerge (together with the hypersensitivity already highlighted by Ward).
A novel coding strategy for block-based compressive sensing named spatially directional predictive coding (SDPC) is proposed, which efficiently utilizes the intrinsic spatial correlation of natural images. At the enco...
详细信息
ISBN:
(纸本)9781479923410
A novel coding strategy for block-based compressive sensing named spatially directional predictive coding (SDPC) is proposed, which efficiently utilizes the intrinsic spatial correlation of natural images. At the encoder, for each block of compressive sensing (CS) measurements, the optimal prediction is selected from a set of prediction candidates that are generated by four designed directional predictive modes. Then, the resulting residual is processed by scalar quantization (SQ). At the decoder, the same prediction is added onto the de-quantized residuals to produce the quantized CS measurements, which is exploited for CS reconstruction. Experimental results substantiate significant improvements achieved by SDPC-plus-SQ in rate distortion performance as compared with SQ alone and DPCM-plus-SQ.
This paper concerns how extant theorists of predictive coding conceptualize and explain possible instances of cognitive penetration. Section 1 offers brief clarification of the predictive coding framework and of cogni...
详细信息
This paper concerns how extant theorists of predictive coding conceptualize and explain possible instances of cognitive penetration. Section 1 offers brief clarification of the predictive coding framework and of cognitive penetration. Section 2 develops more precise ways that the predictive coding framework can explain genuine top-down causal effects on perceptual experience. Section 3 develops these insights further with an eye towards tracking one extant criterion for cognitive penetration, namely, that the relevant cognitive effects on perception must be sufficiently direct. In Section 4, we analyze and criticize a claim made by some theorists of predictive coding, namely, that (interesting) instances of cognitive penetration tend to occur in perceptual circumstances involving substantial noise or uncertainty. We argue that, when applied, the claim fails to explain (or perhaps even be consistent with) a large range of important and uncontroversially interesting possible cases of cognitive penetration. We conclude with a general speculation about how the recent work on the predictive mind may influence the current dialectic concerning top-down effects on perception. (C) 2016 Elsevier Inc. All rights reserved.
We introduce two new source coding problems: robust sequential coding and robust predictive coding. For the Gauss-Markov source model, we characterize certain supporting hyperplanes of the rate region of these two cod...
详细信息
ISBN:
(纸本)9781467325790
We introduce two new source coding problems: robust sequential coding and robust predictive coding. For the Gauss-Markov source model, we characterize certain supporting hyperplanes of the rate region of these two coding problems. Our investigation also reveals a class of extremal inequalities and minimax theorems.
This paper addresses the problem of robust communication of predictively encoded video in a joint source-channel setting. Specifically, the problem of predictive mismatch, where there is a drift between the state of t...
详细信息
ISBN:
(纸本)0769518966
This paper addresses the problem of robust communication of predictively encoded video in a joint source-channel setting. Specifically, the problem of predictive mismatch, where there is a drift between the state of the encoder and the decoder is addressed as a variant of the Wyner-Ziv problem. We propose a video encoding algorithm based on the H.26L video codec which prevents the propagation of error in predictively encoded video in the event of predictive mismatch (or drift) between the encoder and the decoder. One of the main advantages of the proposed approach is that there is minimal loss in performance over the standard H.26L encoder during error-free transmission, while simultaneously allowing error recovery in the event of errors. Using turbo codes as coset codes, we evaluate the performance of the proposed codec and demonstrate the efficacy of the proposed framework. The performance of the proposed approach can only improve with the use of superior coset codes.
暂无评论