The theory of reinforcement learning provides a normative account, deeply rooted in psychological and neuroscientific perspectives on animal behaviour, of how agents may optimize their control of an environment. To use reinforcement learning successfully in situations approaching real-world complexity, however, agents are confronted with a difficult task: they must derive efficient representations of the environment from high-dimensional sensory inputs, and use these to generalize past experience to new situations. Remarkably, humans and other animals seem to solve this problem through a harmonious combination of reinforcement learning and hierarchical sensory processing systems, the former evidenced by a wealth of neural data revealing notable parallels between the phasic signals emitted by dopaminergic neurons and temporal difference reinforcement learning algorithms. While reinforcement learning agents have achieved some successes in a variety of domains, their applicability has previously been limited to domains in which useful features can be handcrafted, or to domains with fully observed, low-dimensional state spaces. Here we use recent advances in training deep neural networks to develop a novel artificial agent, termed a deep Q-network, that can learn successful policies directly from high-dimensional sensory inputs using end-to-end reinforcement learning. We tested this agent on the challenging domain of classic Atari 2600 games. We demonstrate that the deep Q-network agent, receiving only the pixels and the game score as inputs, was able to surpass the performance of all previous algorithms and achieve a level comparable to that of a professional human games tester across a set of 49 games, using the same algorithm, network architecture and hyperparameters. This work bridges the divide between high-dimensional sensory inputs and actions, resulting in the first artificial agent that is capable of learning to excel at a diverse array of challenging tasks.

/pdf/human-level-control-through-deep-reinforcement-learning-40h0gx9pjx.pdf

Human-level control through deep reinforcement learning

Statistical method for testing the neutral mutation hypothesis by DNA polymorphism.

Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Information processing in the cerebral cortex involves interactions among distributed areas. Anatomical connectivity suggests that certain areas form local hierarchical relations such as within the visual system. Other connectivity patterns, particularly among association areas, suggest the presence of large-scale circuits without clear hierarchical relations. In this study the organization of networks in the human cerebrum was explored using resting-state functional connectivity MRI. Data from 1,000 subjects were registered using surface-based alignment. A clustering approach was employed to identify and replicate networks of functionally coupled regions across the cerebral cortex. The results revealed local networks confined to sensory and motor cortices as well as distributed networks of association regions. Within the sensory and motor cortices, functional connectivity followed topographic representations across adjacent areas. In association cortex, the connectivity patterns often showed abrupt transitions between network boundaries. Focused analyses were performed to better understand properties of network connectivity. A canonical sensory-motor pathway involving primary visual area, putative middle temporal area complex (MT+), lateral intraparietal area, and frontal eye field was analyzed to explore how interactions might arise within and between networks. Results showed that adjacent regions of the MT+ complex demonstrate differential connectivity consistent with a hierarchical pathway that spans networks. The functional connectivity of parietal and prefrontal association cortices was next explored. Distinct connectivity profiles of neighboring regions suggest they participate in distributed networks that, while showing evidence for interactions, are embedded within largely parallel, interdigitated circuits. We conclude by discussing the organization of these large-scale cerebral networks in relation to monkey anatomy and their potential evolutionary expansion in humans to support cognition.

The organization of the human cerebral cortex estimated by intrinsic functional connectivity

http://science.sciencemag.org/content/sci/201/4359/899.2.full.pdf

Phenomenology of perception.

The study of decision making spans such varied fields as neuroscience, psychology, economics, statistics, political science, and computer science. Despite this diversity of applications, most decisions share common elements including deliberation and commitment. Here we evaluate recent progress in understanding how these basic elements of decision formation are implemented in the brain. We focus on simple decisions that can be studied in the laboratory but emphasize general principles likely to extend to other settings.

/pdf/the-neural-basis-of-decision-making-2p5agsty56.pdf

The Neural Basis of Decision Making

Relationships between Pupil Diameter and Neuronal Activity in the Locus Coeruleus, Colliculi, and Cingulate Cortex.

http://www.liralab.it/teaching/SINA%5F08%2D09/papers/gold&shadlen2.pdf

Neural computations that underlie decisions about sensory stimuli

Behaviour often depends on the ability to make categorical judgements about sensory information acquired over time. Such judgements require a comparison of the evidence favouring the alternatives1,2,3,4, but how the brain forms these comparisons is unknown. Here we show that in a visual discrimination task, the accumulating balance of sensory evidence favouring one interpretation over another is evident in the neural circuits that generate the behavioural response. We trained monkeys to make a direction judgement about dynamic random-dot motion5 and to indicate their judgement with an eye movement to a visual target. We interrupted motion viewing with electrical microstimulation of the frontal eye field and analysed the resulting, evoked eye movements for evidence of ongoing activity associated with the oculomotor response6,7,8,9,10. Evoked eye movements deviated in the direction of the monkey's judgement. The magnitude of the deviation depended on motion strength and viewing time. The oculomotor signals responsible for these deviations reflected the accumulated motion information that informed the monkey's choices on the discrimination task. Thus, for this task, decision formation and motor preparation appear to share a common level of neural organization.

Representation of a perceptual decision in developing oculomotor commands.

The ability to make inferences about the current state of a dynamic process requires ongoing assessments of the stability and reliability of data generated by that process. Here the authors report that these assessments were reflected in pupil diameter changes, suggesting that pupil-linked arousal systems can help regulate the influence of incoming data on existing beliefs.

Joshua I. Gold

Papers

The Neural Basis of Decision Making

Relationships between Pupil Diameter and Neuronal Activity in the Locus Coeruleus, Colliculi, and Cingulate Cortex.

Neural computations that underlie decisions about sensory stimuli

Representation of a perceptual decision in developing oculomotor commands.

Rational regulation of learning dynamics by pupil-linked arousal systems