Psychiatry: insights into depression through normative decision-making models (2009)
Quentin Jm Huys, Joshua T Vogelstein, Peter Dayan
Decision making lies at the very heart of many psychiatric diseases. It is also a central theoretical concern in a wide variety of fields and has undergone detailed, in-depth, analyses. We take as an...
Estimation of returns over time, the focus of temporal difference (TD) algorithms, imposes particular constraints on good function approxi-mators or representations. Appropriate generalization...
Behavioral/Systems/Cognitive Human Pavlovian–Instrumental Transfer (2009)
Deborah Talmi, Ben Seymour, Peter Dayan, Raymond J. Dolan
The vigor with which a participant performs actions that produce valuable outcomes is subject to a complex set of motivational influences. Many of these are believed to involve the amygdala and the...
Communicated by Steve Nowlan Competition and Multiple Cause Models (2009)
If different causes can interact on any occasion to generate a set of patterns, then systems modeling the generation have to model the in-teraction too. We discuss a way of combining multiple causes...
LETTER Communicated by Lawrence Saul Recurrent Sampling Models for the Helmholtz Machine (2009)
Many recent analysis-by-synthesis density estimation models of cortical learning and processing have made the crucial simplifying assumption that units within a single layer are mutually independent...
A recent article (Stanton and Sejnowski 1989) on long-term synaptic depression in the hippocampus has reopened the issue of the com-putational efficiency of particular synaptic learning rules (Hebb...
The elastic net, which has been used to produce accounts of the forma-tion of topology-preserving maps and ocular dominance stripes (OD), embodies a nearest neighbor topology. A Hebbian account of OD...
Peter Dayan, Terrence J. Sejnowski
(Kohonen 1987; Hinton and Anderson 1981) and are also biologically
Communicated by Robert Jacobs Factor Analysis Using Delta-Rule Wake-Sleep Learning (2009)
We describe a linear network that models correlations between real-valued visible variables using one or more real-valued hidden variables—a factor analysis model. This model can be seen as a...
A Neural Substrate of Prediction and Reward (2009)
M. Romanski, J. E. Ledoux, J. Neurosci, J. L. Armony, J. D. Cohen, ...
The capacity to predict future events permits a creature to detect, model, and manipulate the causal structure of its interactions with its environment. Behavioral experiments suggest that learning...
The Role of Value Systems in Decision Making (2009)
Edited Christoph Engel, Wolf Singer, Peter Dayan
Values, rewards, and costs play a central role in economic, statistical, and psychological notions of decision making. They also have surprisingly direct neural realizations. This chapter discusses...
Hong Xu, Peter Dayan, Richard M. Lipkin, Ning Qian
Adaptation is ubiquitous in sensory processing. Although sensory processing is hierarchical, with neurons at higher levels exhibiting greater degrees of tuning complexity and invariance than those at...
Probabilistic Interpretation of Population Codes Communicated by Terrence Sanger (2009)
Richard S. Zemel, Peter Dayan, Alexandre Pouget
We present a general encoding-decoding framework for interpreting the activity of a population of units. A standard population code interpretation method, the Poisson model, starts from a description...
LETTER Communicated by Sidney Lehky A Hierarchical Model of Binocular Rivalry (2009)
Binocular rivalry is the alternating percept that can result when the two eyes see different scenes. Recent psychophysical evidence supports the notion that some aspects of binocular rivalry bear...
Serotonin, Inhibition, and Negative Mood (2009)
Pavlovian predictions of future aversive outcomes lead to behavioral inhibition, suppression, and withdrawal. There is considerable evidence for the involvement of serotonin in both the learning of...
Reconstructing Woman: From Fiction to Reality in the Nineteenth-Century French Novel (review) (2009)
French Studies: A Quarterly Review - Volume 63, Number 2, April 2009
Behavioral/Systems/Cognitive Differential Encoding of Losses and Gains in the (2009)
Human Striatum, Ben Seymour, Nathaniel Daw, Peter Dayan, Tania Singer, Ray Dolan
Studies on human monetary prediction and decision making emphasize the role of the striatum in encoding prediction errors for financial reward. However, less is known about how the brain encodes...
DOI 10.1007/s10827-005-5705-x (2009)
J Comput Neurosci, Aaron J. Gruber, Peter Dayan, Boris S. Gutkin, Sara A. Solla, A. J. Gruber, ...
Dopamine modulation in the basal ganglia locks the gate to working memory
Submitted to NIPS 2000. Competition and Arbors in Ocular Dominance (2009)
Hebbian and competitive Hebbian algorithms are almost ubiquitous in modeling pattern formation in cortical development. We analyse in theoretical detail a particular model (adapted from Piepenbrock...
Gatsby Computational Neuroscience Unit (2009)
Quantitative data on the speed with which animals acquire behavioral responses during classical conditioning experiments should provide strong constraints on models of learning. However, most models...
LETTER Communicated by Pawan Sinha Images, Frames, and Connectionist Hierarchies (2009)
The representation of hierarchically structured knowledge in systems using distributed patterns of activity is an abiding concern for the connectionist solution of cognitively rich problems. Here, we...
A Bayesian model predicts the response of axons to molecular gradients (2009)
Mortimer, Duncan, Feldner, Julia, Vaughan, Timothy, Vetter, Irina, Pujic, Zac, Rosoff, William J., ...
Axon guidance by molecular gradients plays a crucial role in wiring up the nervous system. However, the mechanisms axons use to detect gradients are largely unknown. We first develop a Bayesian...
A Bayesian model predicts the response of axons to molecular gradients (2009)
Mortimer, Duncan, Feldner, Julia, Vaughan, Timothy, Vetter, Irina, Pujic, Zac, Rossoff, William J., ...
Axon guidance by molecular gradients plays a crucial role in wiring up the nervous system. However, the mechanisms axons use to detect gradients are largely unknown. We first develop a Bayesian...
A Bayesian model predicts the response of axons to molecular gradients (2009)
Mortimer, Duncan, Feldner, Julia, Vaughan, Timothy, Vetter, Irina, Pujic, Zac, Rossoff, William J., ...
Axon guidance by molecular gradients plays a crucial role in wiring up the nervous system. However, the mechanisms axons use to detect gradients are largely unknown. We first develop a Bayesian...
Gatsby Computational Neuroscience Unit, University (2008)
The involvement of recurrent connections in area CA3 in establishing the properties of place fields: A model.
Arbitrary elastic topologies and ocular dominance (2008)
The elastic net, which has been used to produce accounts of the formation of topology preserving mapsand ocular dominance columns (OD), embodies a nearest neighbour topology. A Hebbian account of OD...
A Neural Substrate of Prediction and Reward (2008)
M. Romanski, J. E. Ledoux, J. Neurosci, J. L. Armony, J. D. Cohen, ...
The capacity to predict future events permits a creature to detect, model, and manipulate the causal structure of its interactions with its environment. Behavioral experiments suggest that learning...
Dorsal Striatum in Instrumental (2008)
R Eports, M. Becker-hapak, S. S. Mcallister, S. F. Dowdy, H. P. Grill, J. L. Zweier, ...
24. We thank D. Eliezer and T. McGraw for helpful discussions, K. D’Amico and L. Tong for help with data collection at the Advanced Photon Source, and D. Landry for help with the surface plasmon...
Abstract Conditioning experiments probe the ways that animals make pre-dictions about rewards and punishments and use those predictions to control their behavior. One standard model of condition-ing...
Replay, Repair and Consolidation (2008)
A standard view of memory consolidation is that episodes are stored temporarily in the hippocampus, and are transferred to the neocortex through replay. Various recent experimental challenges to the...
Richard S. Zemel, Rama Natarajan, Peter Dayan
As animals interact with their environments, they must constantly update estimates about their states. Bayesian models combine prior probabilities, a dynamical model and sensory evidence to update...
A Neural Substrate of Prediction and Reward (2008)
M. Romanski, J. E. Ledoux, J. Neurosci, J. L. Armony, J. D. Cohen, ...
The capacity to predict future events permits a creature to detect, model, and manipulate the causal structure of its interactions with its environment. Behavioral experiments suggest that learning...
Perturbing Hebbian Rules, Peter Dayan, Geoffrey Goodhill
correlational rules for synaptic development in the visual system, and Miller [5, 8] has studied such rules in the case of two populations of fibres (particularly two eyes). Miller’s analysis has...
Hippocampal Contributions to Control: The Third Way (2008)
Recent experimental studies have focused on the specialization of different neural structures for different types of instrumental behavior. Recent theoretical work has provided normative accounts for...
In NIPS 9. An Hierarchical Model of Visual Rivalry (2008)
Binocular rivalry is the alternating percept that can result when the two eyes see different scenes. Recent psychophysical evidence supports an account for one component of binocular rivalry similar...
Position Variance, Recurrence and Perceptual Learning (2008)
Stimulus arrays are inevitably presented at different positions on the retina in visual tasks, even those that nominally require fixation. In particular, this applies to many perceptual learning...
Peter Dayan, L. F. Abbott, Dayan Peter, Systems Peter Dayan, L. F. Abbott
All rights reserved. No part of this book may be reproduced in any form by any electronic or mechanical means (including photocopying, recording, or information storage and retrieval) without...
2 Pattern formation and cortical maps (2008)
6 The response selectivities of neurons in adult primary sensory cortices depend on intricate patterns of synaptic connections; these 7 selectivities are arranged over cortex in equally rich fashion....
Abstract In memory consolidation, declarative memories which initially require the hippocampus for their recall, ultimately become independent of it. Consolidation has been the focus of numerous...
Hippocampal Contributions to Control: The Third Way (2008)
Recent experimental studies have focused on the specialization of different neural structures for different types of instrumental behavior. Recent theoretical work has provided normative accounts for...
Table of Contents Reinforcement Learning (2008)
A Formal framework for learning from reinforcement – Markov decision problems – Optimization of long term return
Yael Niv, Daphna Joel, Peter Dayan
Understanding the effects of motivation on instrumental action selection, and specifically on its two main forms, goal-directed and habitual control, is fundamental to the study of decision making....
Odelia Schwartz, Terrence J. Sejnowski, Peter Dayan
Gaussian scale mixture models offer a top-down description of signal generation that captures key bottom-up statistical characteristics of filter responses to images. However, the pattern of...
French Symbolist Poetry and the Idea of Music (review) (2008)
French Studies: A Quarterly Review - Volume 62, Number 3, July 2008
We have calculated analytical expressions for how the bias and variance of the estimators provided by various temporal di erence value estimation algorithms change with o ine updates over trials in...
Peter Dayan, Bernard W. Balleine
There is substantial evidence that dopamine is in-volved in reward learning and appetitive conditioning. However, the major reinforcement learning-based theoretical models of classical conditioning...
In NIPS 11. Computational Differences between Asymmetrical and Symmetrical Networks (2008)
Symmetrically connected recurrent networks have recently been used as models of a host of neural computations. However, because of the separation between excitation and inhibition, biological neural...
Richard S. Zemel, Rama Natarajan, Peter Dayan
As animals interact with their environments, they must constantly update estimates about their states. Bayesian models combine prior probabilities, a dynamical model and sensory evidence to update...
Unsupervised learning is largely concerned with finding structure among sets of input patterns such as visual scenes. One important example of structure comes in cases that the input patterns are...
Conditioning experiments probe the ways that animals make predictions about rewards and punishments and use those predictions to control their behavior. One standard model of conditioning paradigms...
A simple algorithm that discovers e cient perceptual codes (2008)
In M. Jenkin, Biological Mechanisms, Brendan J. Frey, Peter Dayan
We describe the \wake-sleep " algorithm that allows a multilayer, unsupervised, neural network to build a hierarchy ofrepresentations of sensory input. The network has bottom-up \recognition...
Rate- and Phase-coded Autoassociative Memory (2008)
Areas of the brain involved in various forms of memory exhibit patterns of neural activity quite unlike those in canonical computational models. We show how to use well-founded Bayesian probabilistic...
Bilinearity, rules, and prefrontal cortex (2008)
Humans can be instructed verbally to perform computationally complex cognitive tasks; their performance then improves relatively slowly over the course of practice. Many skills underlie these...
Cosyne 2007 Thursday evening, Poster I-2 Bayesian Models of Dynamic Attentional Selection (2008)
Angela J. Yu, Peter Dayan, Jonathan D. Cohen
Selection amongst potentially conflicting inputs is a critical facet of many decision making tasks. According to Bayesian optimality principles, the attentional suppression of irrelevant inputs and...
Unsupervised learning studies how systems can learn to represent particular input pat-terns in a way that reflects the statistical structure of the overall collection of input pat-terns. By contrast...
Odelia Schwartz, Terrence J. Sejnowski, Peter Dayan
In the analysis of natural images, Gaussian scale mixtures (GSM) have been used to account for the statistics of filter responses, and to inspire hierarchical cortical representational learning...
Peter Dayan, Geoffrey E Hinton
One way to speed up reinforcement learning is to enable learning to happen simultaneously at multiple resolutions in space and time. This paper shows how to create a É-learning managerial hierarchy...
1 Gaussian Information Bottleneck Information Bottleneck for Gaussian Variables (2008)
Amir Globerson, Naftali Tishby, Yair Weiss, Peter Dayan
The problem of extracting the relevant aspects of data was previously addressed through the information bottleneck (IB) method, through (soft) clustering one variable while preserving information...
Gatsby Computational Neuroscience Unit, UCL (2008)
Explaining away has mostly been considered in terms of inference of states in belief networks. We show how it can also arise in a Bayesian context in inference about the weights governing...
Acetylcholine, Norepinephrine, and Spatial Attention (2008)
Despite strong implication of the neuromodulators acetylcholine (ACh) and norepinephrine (NE) in cognitive tasks, there is little consensus on their computational functions. We propose that they...
Tree-Structured Neural Decoding (2008)
We propose adaptive testing as a general mechanism for extracting information about stimuli from spike trains. Each test or question corresponds to choosing a neuron and a time interval and checking...
Running Head: ACQUISITION AND EXTINCTION IN AUTOSHAPING Acknowledgements (2008)
are most grateful to Randy Gallistel and John Gibbon for freely sharing, prior to publica-tion, their many ideas about timing and conditioning. We are also very grateful to Nathaniel Daw for...
Peter Dayan, Geoffrey E Hinton
One way to speed up reinforcement learning is to enable learning to happen simultaneously at multiple resolutions in space and time. This paper shows how to create a É-learning managerial hierarchy...
Probabilistic Interpretation of Population Codes Communicated by Terrence Sanger (2008)
Richard S. Zemel, Peter Dayan, Alexandre Pouget
We present a general encoding-decoding framework for interpreting the activity of a population of units. A standard population code interpretation method, the Poisson model, starts from a description...
Richard S. Zemel, Rama Natarajan, Peter Dayan
As animals interact with their environments, they must constantly update estimates about their states. Bayesian models combine prior probabilities, a dynamical model and sensory evidence to update...
Recurrent network models of area CA3 in the hippocampus capture faithfully many of the properties of place cells. However, they seem ill suited to explaining the substantial experimental data on...
Gatsby Computational Neuroscience Unit (2008)
Quantitative data on the speed with which animals acquire behavioral responses during classical conditioning experiments should provide strong constraints on models of learning. However, most models...
D. J. Foster, Peter Dayan, Centre For Neuroscience, Crichton Street
There is an apparent discrepancy in the rodent hippocampal literature, between the putative involvement of hippocampal principal neurons in navigation, and the limited navigational correlates of...
Submitted to NIPS 2000. Explaining Away in Weight Space (2008)
Explaining away has mostly been considered in terms of inference of states in belief networks. We show how it can also arise in a Bayesian context in inference about the weights governing...
Richard S. Zemel, Rama Natarajan, Peter Dayan
As animals interact with their environments, they must constantly update estimates about their states. Bayesian models combine prior probabilities, a dynamical model and sensory evidence to perform a...
Odelia Schwartz, Terrence J. Sejnowski, Peter Dayan
The misjudgement of tilt in images lies at the heart of entertaining visual illusions and rigorous perceptual psychophysics. A wealth of findings has attracted many mechanistic models, but few clear...
Norepinephrine and Neural Interrupts (2008)
Experimental data indicate that norepinephrine is critically involved in aspects of vigilance and attention. Previously, we considered the function of this neuromodulatory system on a time scale of...
Replay, Repair and Consolidation (2008)
A standard view of memory consolidation is that episodes are stored temporarily in the hippocampus, and are transferred to the neocortex through replay. Various recent experimental challenges to the...
A simple algorithm that discovers e cient perceptual codes (2008)
In M. Jenkin, Biological Mechanisms, Brendan J. Frey, Peter Dayan
We describe the \wake-sleep " algorithm that allows a multilayer, unsupervised, neural network to build a hierarchy of representations of sensory input. The network has bottom-up...
Michael Spratling Michael, Peter Dayan
In order to perform object recognition it is necessary to learn representations of the underlying components of images. Such components correspond to objects, object-parts, or features. Nonnegative...
Serotonin, Inhibition, and Negative Mood (2008)
Pavlovian predictions of future aversive outcomes lead to behavioral inhibition, suppression, and withdrawal. There is considerable evidence for the involvement of serotonin in both the learning of...
L'Œil de Platon et le regard romantique (review) (2008)
French Studies: A Quarterly Review - Volume 61, Number 4, October 2007
Encoding and Decoding Spikes for Dynamic Stimuli (2008)
Rama Natarajan, Peter Dayan, Richard S. Zemel
Naturally occurring sensory stimuli are dynamic. In this article, we consider how spiking neural populations might transmit information about continuous dynamic stimulus variables. The combination of...
Running head: Feudal Q-Learning (2007)
Key words: reinforcement learning, dynamic programming, Q-learning hierarchies One popular way of exorcising the daemon of dimensionality in dynamic programming is to consider spatial and temporal...
Maximilian Riesenhuber, Peter Dayan
We present a connectionist method for representing images that explicitly addresses their hierarchical nature. It blends data from neuroscience about whole-object viewpoint sensitive cells in...
the properties of place fields (2007)
Szabolcs Kali, Szabolcs Kali, Peter Dayan, Peter Dayan
The involvement of recurrent
Recurrent Sampling Models (2007)
. Hierarchical probabilistic synthesis and analysis models have recently been suggested as architectures for performing density estimation. Strict hierarchies makes it easy to evaluate generative or...
Yoram Singer, Prof Naftali Tishby, Peter Dayan, Shlomo Dubnov, Shai Fine, Yoav Freund, ...
2 1 Introduction 4 2 Dynamical Encoding of Cursive Handwriting 14 2.1 Introduction : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 14 2.2 The Cycloidal Model : : : :...
Foraging Through Prediction (2007)
Peter Dayan, P Read Montague, Terrence J Sejnowski, P Read, Montague Terrence, J Sejnowski
To survive, an animal must use sensory events to predict the presence of mates, food, danger, and various other stimuli that are important for its survival and procreation. Although reliable...
Neurobiological Modeling: Squeezing Top Down to Meet Bottom Up (2007)
INTRODUCTION A cartoon description of the goals of cognitive science and neuroscience might read respectively "how the mind works" and "how the brain works". In this caricature,...
Tree-Structured Neural Decoding (2007)
We propose adaptive testing as a general mechanism for extracting information about stimuli from spike trains. Each test or question corresponds to choosing a neuron and a time interval and checking...
Gatsby Unit Multiplicative Modulation of (2007)
Bump Attractors, Bump Attractors, Maneesh Sahani, Maneesh Sahani, Peter Dayan, Peter Dayan
A number of electrophysiological studies in visual and visuo-motor cortices have shown that the tuning curves of cells to visual stimulus parameters may be multiplicatively modulated by extra-retinal...
Substantial data support a temporal difference (TD) model of dopamine (DA) neuron activity in which the cells provide a global error signal for reinforcement learning. However, in certain...
Peter Dayan, Queen Square London
The standard reinforcement learning view of the involvement of neuromodulatory systems in instrumental conditioning includes a rather straightforward conception of motivation as prediction of sum...
Position Variance, Recurrence and Perceptual Learning (2007)
Stimulus arrays are inevitably presented at different positions on the retina in visual tasks, even those that nominally require fixation. In particular, this applies to many perceptual learning...
Acetylcholine (ACh) has been implicated in a wide variety of tasks involving attentional processes and plasticity. Following extensive animal studies, it has previously been suggested that ACh...
Gatsby Computational Neuroscience Unit, UCL (2007)
Explaining away has mostly been considered in terms of inference of states in belief networks. We show how it can also arise in a Bayesian context in inference about the weights governing...
Long Term Potentiation, Navigation & Dynamic Progamming (2007)
Blum and Abbott (1995) recently proposed an algorithm for learned navigation that is based on Hebbian changes to adaptive connections between place cells in the hippocampus. This paper suggests using...
Section: Applications Preference: Oral Observation Biases in Diagnostic Inference (2007)
In real-world statistical inference problems, such as the diagnosis of diseases from the results of medical tests and procedures, there are systematic observation biases as to which test results are...
Manuscript: 1646 Recurrent Sampling Models for the Helmholtz Machine (2007)
Many recent analysis-by-synthesis density estimation models of cortical learning and processing have made the crucial simplifying assumption that units within a single layer are mutually independent...
Gatsby Computational Neuroscience Unit, UCL (2007)
Explaining away has mostly been considered in terms of inference of states in belief networks. We show how it can also arise in a Bayesian context in inference about the weights governing...
The method of temporal differences (TD) is one way of making consistent predictions about the future. This paper uses some analysis of Watkins [19] to extend a convergence theorem due to Sutton [17]...
Serotonin, Inhibition and Negative Mood (2007)
Pavlovian predictions of future aversive outcomes lead to behavioral inhibition, suppression and withdrawal. There is considerable evidence for the involvement of serotonin in both the learning of...
Debussy and the Fragment (review) (2007)
Nineteenth Century French Studies - Volume 35, Number 3&4, Spring-Summer 2007
Mallarmé and Debussy: Unheard Music, Unseen Text (review) (2007)
French Studies: A Quarterly Review - Volume 60, Number 3, July 2006
The selectivities of neurons in primary visual cortex are often considered to be adapted to the statistics of natural images. Accordingly, simple cell-like tuning emerges when unsupervised learning...
phase and oscillatory hippocampal recall (2007)
Many neural areas, notably, the hippocampus, show structured, dynamical, population behavior such as coordinated oscillations. It has long been observed that such oscillations provide a substrate for...
Richard S. Zemel, Rama Natarajan, Peter Dayan
Uncertainty coming from the noise in its neurons and the ill-posed nature of many tasks plagues neural computations. Maybe surprisingly, many studies show that the brain manipulates these forms of...
phase and oscillatory hippocampal recall (2007)
Many neural areas, notably, the hippocampus, show structured, dynamical, population behavior such as coordinated oscillations. It has long been observed that such oscillations provide a substrate for...
Quentin Jm Huys, Richard S Zemel, Rama Natarajan, Peter Dayan
Uncertainty arises in neural computations from noisy processing elements and the formally ill-posed nature of many tasks. Taking appropriate decisions requires that uncertainty be represented and...
Semi-rational Models of Conditioning: The Case of Trial Order (2007)
Nathaniel D. Daw, Aaron C. Courville, Peter Dayan
Bayesian treatments of animal conditioning start from a generative model that specifies precisely a set of assumptions about the structure of the learning task. Optimal rules for learning are direct...
Semi-rational Models of Conditioning: The Case of Trial Order (2007)
Nathaniel D. Daw, Aaron C. Courville, Peter Dayan
Bayesian treatments of animal conditioning start from a generative model that specifies precisely a set of assumptions about the structure of the learning task. Optimal rules for learning are direct...
Yael Niv, Nathaniel D. Daw, Daphna Joel, Peter Dayan, Y. Niv, N. D. Daw, ...
Rationale Dopamine neurotransmission has long been known to exert a powerful influence over the vigor, strength, or rate of responding. However, there exists no clear understanding of the...
Distance Patterns in Structural Similarity (2006)
Similarity of edge labeled graphs is considered in the sense of minimum squared distance between corresponding values. Vertex correspondences are established by isomorphisms if both graphs are of...
Phasic norepinephrine: A neural interrupt signal for unexpected events. Network 17 (2006)
Extensive animal studies indicate that the neuromodulator norepinephrine plays an important role in specific aspects of vigilance, attention and learning, putatively serving as a neural interrupt or...
Dopamine, uncertainty and TD learning (2005)
Niv, Yael, Duff, Michael O, Dayan, Peter
Abstract Substantial evidence suggests that the phasic activities of dopaminergic neurons in the primate midbrain represent a temporal difference (TD) error in predictions of future reward, with...
Nathaniel D. Daw, Yael Niv, Peter Dayan
The basal ganglia are widely believed to be involved in the learned selection of actions. Building on this idea, reinforcement learning (RL) theories of optimal control have had some success in...
Inference, attention, and decision in a Bayesian neural architecture (2005)
We study the synthesis of neural coding, selective attention and perceptual decision making. A hierarchical neural architecture is proposed, which implements Bayesian integration of noisy sensory...
Estimation of non-normalized statistical models by score matching (2005)
One often wants to estimate statistical models where the probability density function is known only up to a multiplicative normalization constant. Typically, one then has to resort to Markov Chain...
Behavioral and Brain Functions 2005, 1:6 doi:10.1186/1744-9081-1-6 (2005)
Yael Niv, Michael O. Duff, Peter Dayan, Yael Niv, Michael O. Duff, Peter Dayan
PDF corresponds to the article as it appeared upon acceptance. The fully-formatted PDF version will become available shortly after the date of publication, from the URL listed below. Dopamine,...
Differential priors for elastic nets (2005)
Peter Dayan, Geoffrey J. Goodhill
Abstract. The elastic net and related algorithms, such as generative topographic mapping, are key methods for discretized dimension-reduction problems. At their heart are priors that specify the...
A review of "On Intelligence", a book exploring brain function by entrepreneur Jeff Hawkins and science writer Sandra Blakeslee.
Non-negative matrix factorization with sparseness constraints (2004)
Non-negative matrix factorization (NMF) is a recently developed technique for finding parts-based, linear representations of non-negative data. Although it has successfully been applied in several...
Expected and unexpected uncertainty: ACh and NE in the neocortex (2003)
Inference and adaptation in noisy and changing, rich sensory environments are rife with a variety of specific sorts of variability. Experimental and theoretical studies suggest that these different...
Temporal difference models and reward-related learning in the human brain (2003)
Peter Dayan, Karl Friston, Hugo Critchley, London Wcn Bg
John P. O’Doherty, V(tUCS) and V(tUCS � 1) generates a positive prediction error that, in the simplest form of TD learning, is used to increment the value at time tUCS – 1 (in proportion to
Learning Behavior-Selection by Emotions and Cognition in a Multi-Goal Robot Task (2003)
Sandra Clara Gadanho, Peter Dayan
The existence of emotion and cognition as two interacting systems, both with important roles in decision-making, has been recently advocated by neurophysiological research (LeDoux, 1998, Damasio,...
Expected and unexpected uncertainty: ACh and NE in the neocortex (2003)
Inference and adaptation in noisy and changing, rich sensory environments are rife with a variety of specific sorts of variability. Experimental and theoretical studies suggest that these different...
Information bottleneck for gaussian variables (2003)
Amir Globerson, Naftali Tishby, Yair Weiss, Peter Dayan
The problem of extracting the relevant aspects of data was previously addressed through the information bottleneck (IB) method, through (soft) clustering one variable while preserving information...
Dopamine Modulation in a Basal Ganglio-Cortical Network Implements Saliency-Based Gating of (2003)
Working Memory Aaron, Aaron J. Gruber, Peter Dayan, Boris S. Gutkin, Sara A. Solla
Dopamine exerts two classes of effect on the sustained neural activity in prefrontal cortex that underlies working memory. Direct release in the cortex increases the contrast of prefrontal neurons,...
Learning Behavior-Selection by Emotions and Cognition in a Multi-Goal Robot Task (2003)
Sandra Clara Gadanho, Peter Dayan
The existence of emotion and cognition as two interacting systems, both with important roles in decision-making, has been recently advocated by neurophysiological research (LeDoux, 1998, Damasio,...
La musique et les lettres chez Barthes (2003)
Dans ses écrits sur la musique, dès l'époque des Mythologies et jusqu'à la fin de sa vie, Barthes répète le geste symboliste, mallarméen, qui consiste à définir la littérature (voire l'art...
Perceptual inference fundamentally involves uncertainty, arising from noise in sensation and the ill-posed nature of many perceptual problems. Accurate perception requires that this uncertainty be...
Acquisition and extinction in autoshaping (2002)
C. R. Gallistel and J. Gibbon (2000) presented quantitative data on the speed with which animals acquire behavioral responses during autoshaping, together with a statistical model of learning...
ACh, uncertainty, and cortical inference (2002)
Acetylcholine (ACh) has been implicated in a wide variety of tasks involving attentional processes and plasticity. Following extensive animal studies, it has previously been suggested that ACh...
Gatsby Computational Neuroscience Unit (2002)
It is a commonplace in statistics that uncertainty about parameters drives learning. Indeed one of the most influential models of behavioural learning has uncertainty at its heart. However, many...
rence learning. What we will see is that although prediction is relatively straightforward at a systems level, it poses some interesting and tricky conceptual, architectural and mechanistic problems...
ACh, uncertainty, and cortical inference (2002)
Acetylcholine (ACh) has been implicated in a wide variety of tasks involving attentional processes and plasticity. Following extensive animal studies, it has previously been suggested that ACh...
2002 Special issue Acetylcholine in cortical inference (2001)
Acetylcholine (ACh) plays an important role in a wide variety of cognitive tasks, such as perception, selective attention, associative learning, and memory. Extensive experimental and theoretical...
2002 Special issue Dopamine: generalization and bonuses (2001)
In the temporal difference model of primate dopamine neurons, their phasic activity reports a prediction error for future reward. This model is supported by a wealth of experimental data. However, in...
Competition and arbors in ocular dominance (2001)
Hebbian and competitive Hebbian algorithms are almost ubiquitous in modeling pattern formation in cortical development. We analyse in theoretical detail a particular model (adapted from Piepenbrock...
A model of hippocampally dependent navigation, using the temporal difference learning rule (2000)
ABSTRACT: This paper presents a model of how hippocampal place cells might be used for spatial navigation in two watermaze tasks: the standard reference memory task and a delayed matching-to-place...
Learning To Evaluate Go Positions Via Temporal Difference Methods (2000)
Nicol N. Schraudolph, Peter Dayan, Terrence J. Sejnowski
The game of Go has a high branching factor that defeats the tree search approach used in computer chess, and long-range spatiotemporal interactions that make position evaluation extremely difficult....
Strong constraints on the neural mechanisms underlying the formation of place fields in the rodent hippocampus come from the systematic changes in spatial activity patterns that are consequent on...
Learning To Evaluate Go Positions Via Temporal Difference Methods (2000)
Nicol N. Schraudolph, Peter Dayan, Terrence J. Sejnowski
The game of Go has a high branching factor that defeats the tree search approach used in computer chess, and long-range spatiotemporal interactions that make position evaluation extremely difficult....
Hippocampally-dependent consolidation in a hierarchical model of neocortex (2000)
In memory consolidation, declarative memories which initially require the hippocampus for their recall, ultimately become independent of it. Consolidation has been the focus of numerous experimental...
Learning To Evaluate Go Positions Via Temporal Difference Methods (2000)
Nicol N. Schraudolph, Peter Dayan, Terrence J. Sejnowski
Introduction 1.1 The Game of Go Go was developed four millennia ago in China; it is one of the oldest and most popular board games in the world. Like chess, it is a deterministic, perfect...
Acquisition in Autoshaping (2000)
Quantitative data on the speed with which animals acquire behavioral responses during classical conditioning experiments should provide strong constraints on models of learning. However, most models...
Models of Hippocampally Dependent Navigation, Using The Temporal Difference Learning Rule (2000)
D. J. Foster, Peter Dayan, Centre For Neuroscience
this paper investigates models of spatial learning in two navigational tasks, combining TD learning with a place cell representation to learn about rewards and coordinates. We pose the question: can...
The effect of correlated variability on the accuracy of a population code (1999)
We study the impact of correlated neuronal firing rate variability on the accuracy with which an encoded quantity can be extracted from a population of neurons. Contrary to a widespread belief,...
Gatsby Computational Neuroscience Unit, London (1999)
David Foster, Peter Dayan, Satinder Singh
Abstract. Solving in an efficient manner many different optimal control tasks within the same underlying environment requires decomposing the environment into its computationally elemental fragments....
Distributional population codes and multiple motion models (1999)
Most theoretical and empirical studies of population codes make the assumption that underlying neuronal activities is a unique and unambiguous value of an encoded quantity. However, population...
Peter Dayan, Maneesh Sahani, Grégoire Deback
Adaptation is a ubiquitous neural and psychological phenomenon, with a wealth of instantiations and implications. Although a basic form of plasticity, it has, bar some notable exceptions, attracted...
Spatial representations in related environments in a recurrent model of area CA3 of the rat (1999)
Recurrent network models of area CA3 in the hippocampus capture faithfully many of the properties of place cells. However, they seem ill suited to explaining the substantial experimental data on...
The Effect Of Correlated Variability On The Accuracy Of A Population Code (1999)
Abbott Volen Center, L. F. Abbott, Peter Dayan
We study the impact of correlated neuronal firing rate variability on the accuracy with which an encoded quantity can be extracted from a population of neurons. Contrary to a widespread belief,...
Curved Gaussian Models with Application to the Modeling of Foreign Exchange Rates (1999)
this paper, we present a simple extension to a class of non-- linear, volume preserving transformations which provides an efficient local description of curvature. The resulting generalized Gaussian...
Spatial Representations in Related Environments in a Recurrent Model of Area CA3 of the Rat (1999)
Szabolcs Kali And, Szabolcs Kali, Peter Dayan
Recurrent network models of area CA3 in the hippocampus capture faithfully many of the properties of place cells. However, they seem ill suited to explaining the substantial experimental data on...
Statistical Models and Sensory Attention (1999)
Physiological investigations into the neural basis of sensory attention have led to puzzling and contradictory results. Attention can seemingly lead to increased, decreased and unchanged neural...
Computational Differences Between Asymmetrical and Symmetrical Networks (1999)
. Symmetrically connected recurrent networks have recently been used as models of a host of neural computations. However, biological neural networks have asymmetrical connections, at the very least...
Distributional Population Codes and Multiple Motion Models (1999)
Most theoretical and empirical studies of population codes make the assumption that underlying neuronal activities is a unique and unambiguous value of an encoded quantity. However, population...
The Effect Of Correlated Variability On The Accuracy Of A Population Code (1999)
We study the impact of correlated neuronal firing rate variability on the accuracy with which an encoded quantity can be extracted from a population of neurons. Contrary to a widespread belief,...
Statistical Models of Conditioning (1999)
Peter Dayan, Cognitive Sciences, Theresa Long
Conditioning experiments probe the ways that animals make predictions about rewards and punishments and use those predictions to control their behavior. One standard model of conditioning paradigms...
Spatial representations in related environments in a recurrent model of area CA3 of the rat (1999)
Recurrent network models of area CA3 in the hippocampus capture faithfully many of the properties of place cells. However, they seem ill suited to explaining the substantial experimental data on...
Curved gaussian models with application to modeling of foreign exchange rates (1999)
Gaussian distributions lie at the heart of popular tools for capturing structure in high dimensional data. Standard techniques employ as models arbitrary linear transformations of spherical...
Manuscript: 1646 Recurrent Sampling Models for the Helmholtz Machine (1999)
Many recent analysis-by-synthesis density estimation models of cortical learning and processing have made the crucial simplifying assumption that units within a single layer are mutually independent...
Peter Dayan, Maneesh Sahani, Grégoire Deback
Adaptation is a ubiquitous neural and psychological phenomenon, with a wealth of instantiations and implications. Although a basic form of plasticity, it has, bar some notable exceptions, attracted...
Gatsby Computational Neuroscience Unit, London �����¢��� � ¢¡���¡�� (1999)
David Foster, Peter Dayan, Satinder Singh
Abstract. Solving in an efficient manner many different optimal control tasks within the same underlying environment requires decomposing the environment into its computationally elemental fragments....
Curved gaussian models with application to modeling of foreign exchange rates (1999)
Gaussian distributions lie at the heart of popular tools for capturing structure in high dimensional data. Standard techniques employ as models arbitrary linear transformations of spherical...
The effect of correlated variability on the accuracy of a population code (1999)
We study the impact of correlated neuronal firing rate variability on the accuracy with which an encoded quantity can be extracted from a population of neurons. Contrary to a widespread belief,...
The Effect of Correlated Variability on the Accuracy of a Population Code (1999)
We study the impact of correlated neuronal firing rate variability on the accuracy with which an encoded quantity can be extracted from a population of neurons. Contrary to widespread belief,...
A hierarchical model of binocular rivalry (1998)
Binocular rivalry is the alternating percept that can result when the two eyes see different scenes. Recent psychophysical evidence supports the no-tion that some aspects of binocular rivalry bear...
Neurobiological modeling: squeezing top down to meet bottom up (1998)
P Read Montague, Peter Dayan, William Bechtel, George Graham
A cartoon description of the goals of cognitive science and neuroscience might read respectively “how the mind works ” and “how the brain works”. In this caricature, there would seem to be...
Bayesian Retrieval in Associative Memories with Storage Errors (1998)
Friedrich T. Sommer, Peter Dayan
It is well known that for finite-sized networks, one-step retrieval in the autoassociative Willshaw net is a suboptimal way to extract the information stored in the synapses. Iterative retrieval...
Analytical mean squared error curves for temporal difference learning (1998)
Satinder Singh, Peter Dayan, G. Barto
Abstract. We provide analytical expressions governing changes to the bias and variance of the lookup table estimators provided by various Monte Carlo and temporal difference value estimation...
Probabilistic interpretation of population codes (1998)
Richard S. Zemel, Peter Dayan, Alexandre Pouget
We present a general encoding-decoding framework for interpreting the activity of a population of units. A standard population code interpretation method, the Poisson model, starts from a description...
Combining probabilistic population codes (1997)
We study the problem of statistically correct inference in networks whose basic representations are population codes. Population codes are ubiquitous in the brain, and involve thesimultaneous...
Recognition in hierarchical models (1997)
Abstract. Various proposals have recently been made which cast cortical processing in terms of hierarchical statistical generative models
Modeling the manifolds of images of handwritten digits (1997)
Geoffrey E. Hinton, Peter Dayan, Michael Revow
description length, density estimation.
Neural models for part-whole hierarchies (1997)
Maximilian Riesenhuber, Peter Dayan
We present a connectionist method for representing images that explicitly addresses their hierarchical nature. It blends data from neuroscience about whole-object viewpoint sensitive cells in...
Factor analysis using delta-rule wake-sleep learning (1997)
We describe a linear network that models correlations between real-valued visible variables using one or more real-valued hidden variables — a factor analysis model. This model can be seen as a...
Modeling the manifolds of images of handwritten digits (1997)
This paper describes two new methods for modelling the manifolds of digitised images of handwritten digits. The models allow a priori information about the structure of the manifolds to be combined...
Factor analysis using delta-rule wake-sleep learning (1997)
We describe a linear network that models correlations between real-valued visible variables using one or more real-valued hidden variables — a factor analysis model. This model can be seen as a...
Neural models for part-whole hierarchies (1997)
Maximilian Riesenhuber, Peter Dayan
We present a connectionist method for representing images that explicitly addresses their hierarchical nature. It blends data from neuroscience about whole-object viewpoint sensitive cells in...
Recognition in hierarchical models (1997)
Abstract. Various proposals have recently been made which cast cortical processing in terms of hierarchical statistical generative models
Combining Probabilistic Population Codes (1997)
We study the problem of statistically correct inference in networks whose basic representations are population codes. Population codes are ubiquitous in the brain, and involve the simultaneous...
Probabilistic Interpretation of Population Codes (1997)
Richard Zemel, Peter Dayan, Alexandre Pouget
We present a general encoding-decoding framework for interpreting the activity of a population of units. A standard population code interpretation method, the Poisson model, starts from a description...
Using EM for reinforcement learning (1997)
Peter Dayan, Geoffrey E Hinton
We discsus Hinton’s (1989) relative payoff procedure (RPP), a static reinforcement learning algorithm whose foundation is not stochastic gradient ascent. We show circumstances under which applying...
Probabilistic Interpretation of Population Codes (1997)
Richard S. Zemel, Peter Dayan, Alexandre Pouget
We present a general encoding-decoding framework for interpreting the activity of a population of units. A standard population code interpretation method, the Poisson model, starts from a description...
Probabilistic Interpretation of Population Codes (1997)
Richard Zemel, Peter Dayan, Alexandre Pouget
We present a theoretical framework for population codes which generalises naturally to the important case where the population provides information about a whole probability distribution over an...
Recurrent Sampling Models for the Helmholtz Machine (1997)
Many recent analysis-by-synthesis density estimation models of cortical learning and processing have made the crucial simplifying assumption that units within a single layer are mutually independent...
Neural Models for Part-Whole Hierarchies (1997)
Maximilian Riesenhuber, Peter Dayan
We present a connectionist method for representing images that explicitly addresses their hierarchical nature. It blends data from neuroscience about whole-object viewpoint sensitive cells in...
A Simple Algorithm That Discovers Efficient Perceptual Codes (1997)
Brendan Frey, Peter Dayan, Geoffrey E. Hinton, In M. Jenkin, Biological Mechanisms
We describe the "wake-sleep" algorithm that allows a multilayer, unsupervised, neural network to build a hierarchy of representations of sensory input. The network has bottom-up...
A Hierarchical Model of Binocular Rivalry (1997)
Binocular rivalry is the alternating percept that can result when the two eyes see different scenes. Recent psychophysical evidence supports the notion that some aspects of binocular rivalry bear...
Modelling the Manifolds of Images of Handwritten Digits (1997)
Geoffrey E. Hinton, Peter Dayan, Michael Revow
This paper describes two new methods for modelling the manifolds of digitised images of handwritten digits. The models allow a priori information about the structure of the manifolds to be combined...
Using em for reinforcement learning (1997)
Peter Dayan, Geoffrey E Hinton
We discsus Hinton’s (1989) relative payoff procedure (RPP), a static reinforcement learning algorithm whose foundation is not stochastic gradient ascent. We show cir-cumstances under which applying...
Combining probabilistic population codes (1997)
zemelOu.arizona.edu We study the problem of statistically correct inference in networks whose basic representations are population codes. Population codes are ubiquitous in the brain, and involve the...
Improving policies without measuring merits (1996)
Performing policy iteration in dynamic programming should only require knowledge of relative rather than absolute measures of the utility of actions-- what Baird (1993) calls the advantages of...
Exploration bonuses and dual control (1996)
Peter Dayan, Terrence J Sejnowski
certainty equivalence Finding the Bayesian balance between exploration and exploitation in adaptive optimal control is in general intractable. This paper shows how to compute suboptimal estimates...
Factor Analysis Using Delta-Rule Wake-Sleep Learning (1996)
this paper, we suggest the statistical technique of factor analysis as an interesting alternative to principal components analysis, and show how to implement it using an algorithm whose demands on...
Does the Wake-sleep Algorithm Produce Good Density Estimators? (1996)
Brendan J. Frey, Geoffrey E. Hinton, Peter Dayan
The wake-sleep algorithm (Hinton, Dayan, Frey and Neal 1995) is a relatively efficient method of fitting a multilayer stochastic generative model to high-dimensional data. In addition to the top-down...
Exploration Bonuses and Dual Control (1996)
Peter Dayan, Terrence Sejnowski
Finding the Bayesian balance between exploration and exploitation in adaptive optimal control is in general intractable. This paper shows how to compute suboptimal estimates based on a certainty...
Varieties of Helmholtz Machine (1996)
Peter Dayan, Geoffrey E. Hinton
The Helmholtz machine is a new unsupervised learning architecture that uses topdown connections to build probability density models of input and and bottom up connections to build inverses to those...
A framework for mesencephalic dopamine systems based on predictive Hebbian learning (1996)
P. Read Montague, Peter Dayan, Terrence J. Sejnowskw
We develop a theoretical framework that shows how mesen-cephalic dopamine systems could distribute to their targets a signal that represents information about future expectations. In particular, we...
Exploration Bonuses and Dual Control. Machine Learning (1996)
Peter Dayan, Terrence Sejnowski, G. Barto
Abstract. Finding the Bayesian balance between exploration and exploitation in adaptive optimal control is in general intractable. This paper shows how to compute suboptimal estimates based on a...
Recognizing handwritten digits using mixtures of linear models (1995)
Geoffrey E Hinton, Michael Revow, Peter Dayan
We construct a mixture of locally linear generative models of a collection of pixel-based images of digits, and use them for recognition. Different models of a given digit are used to capture...
Competition and multiple cause models (1995)
If different causes can interact on any occasion to generate a set of patterns, then systems modelling the generation have to model the interaction too. We discuss a way of combining multiple causes...
Peter Dayan, Geoffrey E Hinton
The Helmholtz machine is a new unsupervised learning architecture that uses topdown and bottom up connections to build probability density models of input and inverses to those models. The wake-sleep...
Competition and Multiple Cause Models (1995)
If different causes can interact on any occasion to generate a set of patterns, then systems modelling the generation have to model the interaction too. We discuss a way of combining multiple causes...
Peter Dayan, Geoffrey Hinton, Radford Neal, Richard Zemel
Discovering the structure inherent in a set of patterns is a fundamental aim of statistical inference or learning. One fruitful approach is to build a parameterised stochastic generative model,...
The Wake-Sleep Algorithm for Unsupervised Neural Networks (1995)
Geoffrey Hinton, Peter Dayan, Brendan J Frey, Radford M Neal
We describe an unsupervised learning algorithm for a multilayer network of stochastic neurons. Bottom-up "recognition" connections convert the input into representations in successive...
Peter Dayan, Geoffrey E Hinton, Radford M Neal, Richard S Zemel
Discovering the structure inherent in a set of patterns is a fundamental aim of statistical inference or learning. One fruitful approach is to build a parameterised stochastic generative model,...
The Wake-Sleep Algorithm for Unsupervised Neural Networks (1995)
Geoffrey Hinton Peter, Peter Dayan, Brendan J Frey, Radford M Neal
An unsupervised learning algorithm for a multilayer network of stochastic neurons is described. Bottom-up "recognition" connections convert the input into representations in successive...
Peter Dayan, Geoffrey E. Hinton, Radford M. Neal, Richard S. Zemel
Discovering the structure inherent in a set of patterns is a fundamental aim of statistical inference or learning. One fruitful approach is to build a parameterised stochastic generative model,...
The Wake-Sleep Algorithm for Unsupervised Neural Networks (1995)
Geoffrey E Hinton, Peter Dayan, Brendan J Frey, Radford M Neal
An unsupervised learning algorithm for a multilayer network of stochastic neurons is described. Bottom-up "recognition" connections convert the input into representations in successive...
Recognizing handwritten digits using mixtures of linear models (1995)
Geoffrey E Hinton, Michael Revow, Peter Dayan
We construct a mixture of locally linear generative models of a collection of pixel-based images of digits, and use them for recognition. Different models of a given digit are used to capture...
Peter Dayan, Geoffrey E Hinton, Radford M Neal, Richard S Zemel
Discovering the structure inherent in a set of patterns is a fundamental aim of statistical inference or learning. One fruitful approach is to build a parameterised stochastic generative model,...
Geoffrey E Hinton, Peter Dayan, Brendan J Frey, Radford M Neal
wake-sleep algorithm for unsupervised
Recognizing Handwritten Digits Using Mixtures of Linear Models (1995)
Geoffrey E Hinton, Michael Revow, Peter Dayan
We construct a mixture of locally linear generative models of a collection of pixel-based images of digits, and use them for recognition.
Peter Dayan, Geoffrey E Hinton, Radford M Neal, Richard S Zemel
Discovering the structure inherent in a set of patterns is a fundamental aim of statistical inference or learning. One fruitful approach is to build a parameterised stochastic generative model,...
UCSD School of Medicine independent study project : 1994, nos. 9-16. (1994)
Mishra, Dev K., Fithian, Donald C., Balen, Paul F., Daniel, Dale M., Rosen, Peter., Clark, Richard., ...
Thesis (M.D.)--University of California, San Diego, 1994.
Temporal difference learning of position evaluation in the game of Go (1994)
Nicol N. Schraudolph, Peter Dayan, Terrence J. Sejnowski
The game of Go has a high branching factor that defeats the tree search approach used in computer chess, and long-range spatiotemporal interactions that make position evaluation extremely difficult....
Computational modelling (1994)
Computational modelling is playing an increasingly accepted
TD(λ) converges with probability 1 (1994)
Peter Dayan, Terrence J Sejnowski
The methods of temporal differences (Samuel, 1959; Sutton 1984, 1988) allow agents to learn accurate predictions about stationary stochastic future outcomes. The learning is effectively stochastic...
Computational modelling (1994)
Computational modelling is playing a more accepted and more important role in neuroscience. It is not a unitary enterprise, though, and the distinction between two different sorts of modelling, one...
TD(λ) Converges with Probability 1 (1994)
Peter Dayan, Terrence J Sejnowski
The methods of temporal differences (Samuel, 1959; Sutton 1984, 1988) allow agents to learn accurate predictions about stationary stochastic future outcomes. The learning is effectively stochastic...
Temporal Difference Learning of Position Evaluation in the Game of Go (1994)
Nicol N. Schraudolph, Peter Dayan, Terrence J. Sejnowski
The game of Go has a high branching factor that defeats the tree search approach used in computer chess, and long-range spatiotemporal interactions that make position evaluation extremely difficult....
Geoffrey E Hinton, Peter Dayan, Brendan J Frey, Radford M Neal
wake-sleep algorithm for unsupervised
Arbitrary elastic topologies and ocular dominance (1993)
The elastic net, which has been used to produce accounts of the formation of topology preserving maps and ocular dominance columns (OD), embodies a nearest neighbour topology. A Hebbian account of OD...
Improving generalisation for temporal difference learning: The successor representation (1993)
Estimation of returns over time, the focus of temporal difference (TD) algorithms, imposes particular constraints on good function approximators or representations. Appropriate generalisation between...
Feudal Reinforcement Learning (1993)
Peter Dayan, Geoffrey E. Hinton
One way to speed up reinforcement learning is to enable learning to happen simultaneously at multiple resolutions in space and time. This paper shows how to create a Q-learning managerial hierarchy...
Improving generalisation for temporal difference learning: The successor representation (1993)
Estimation of returns over time, the focus of temporal difference (TD) algorithms, imposes particular constraints on good function approximators or representations. Appropriate generalisation between...
Perturbing Hebbian Rules (1992)
Perturbing Hebbian Rules, Peter Dayan, Geoffrey Goodhill
Recently Linsker [2] and MacKay and Miller [3, 4] have analysed Hebbian correlational rules for synaptic development in the visual system, and Miller [5, 8] has studied such rules in the case of two...
The convergence of TD(X) for general k (1992)
Abstract. The method of temporal differences (TD) is one way of making consistent predictions about the futgre.
Altered States and Virtual Beliefs (1990)
Functionalism avoids a potentially fatal infinite regress by realising the low level phenomena of mind in mere Turing machines rather than via undischarged homunculae. Adopting a narrow, logical view...
Analytical Mean Squared Error Curves for Temporal Difference Learning (1988)
Satinder Singh, Peter Dayan, G. Barto
We provide analytical expressions governing changes to the bias and variance of the lookup table estimators provided by various Monte Carlo and temporal difference value estimation algorithms with...
Analytical Mean Squared Error Curves in Temporal Difference Learning (1988)
Satinder Singh, Peter Dayan, Cognitive Sciences
We have calculated analytical expressions for how the bias and variance of the estimators provided by various temporal difference value estimation algorithms change with offline updates over trials...
A review of "On Intelligence", a book exploring brain function by entrepreneur Jeff Hawkins and science writer Sandra Blakeslee
Dopamine, uncertainty and TD learning
Niv, Yael, Duff, Michael O, Dayan, Peter
Substantial evidence suggests that the phasic activities of dopaminergic neurons in the primate midbrain represent a temporal difference (TD) error in predictions of future reward, with increases...
A review of "On Intelligence", a book exploring brain function by entrepreneur Jeff Hawkins and science writer Sandra Blakeslee
Dopamine, uncertainty and TD learning
Niv, Yael, Duff, Michael O, Dayan, Peter
Substantial evidence suggests that the phasic activities of dopaminergic neurons in the primate midbrain represent a temporal difference (TD) error in predictions of future reward, with increases...
Serotonin, Inhibition, and Negative Mood
Dayan, Peter, Huys, Quentin J. M
Pavlovian predictions of future aversive outcomes lead to behavioral inhibition, suppression, and withdrawal. There is considerable evidence for the involvement of serotonin in both the learning of...
Bilinearity, Rules, and Prefrontal Cortex
Humans can be instructed verbally to perform computationally complex cognitive tasks; their performance then improves relatively slowly over the course of practice. Many skills underlie these...
Simple Substrates for Complex Cognition
Complex cognitive tasks present a range of computational and algorithmic challenges for neural accounts of both learning and inference. In particular, it is extremely hard to solve them using the...