We consider finite horizon Markov decision processes under performance measures that involve both the mean and the variance of the cumulative reward. We show that either randomized or history-based policies can improve performance. We prove that the complexity of computing a policy that maximizes the mean reward under a variance constraint is NP-hard for some cases, and strongly NP-hard for oth...

We consider finite horizon Markov decision processes under performance measures that involve both the mean and the variance of the cumulative reward. We show that either randomized or history-based policies can improve performance. We prove that the complexity of computing a policy that maximizes the mean reward under a variance constraint is NP-hard for some cases, and strongly NP-hard for oth...

We consider risk minimization problems for Markov decision processes. From a standpoint of making the risk of random reward variable at each time as small as possible, a risk measure is introduced using conditional value-at-risk for random immediate reward variables in Markov decision processes, under whose risk measure criteria the risk-optimal policies are characterized by the optimality equa...

Abnormal reinforcement learning and representations of reward value are present in schizophrenia, and these impairments can manifest as deficits in risk/reward decision making. These abnormalities may be due in part to dopaminergic dysfunction within cortico-limbic-striatal circuitry. Evidence from studies with laboratory animal have revealed that normal DA activity within different nodes of th...

Functional magnetic resonance imaging (fMRI) has proven a useful tool for observing neural BOLD signal changes during complex cognitive and emotional tasks. Yet the meaning and applicability of the fMRI data being gathered is still largely unknown. The brain's reward system underlies the fundamental neural processes of goal evaluation, preference formation, positive motivation, and choice behav...

Aberrant rule- and reward-based processes underpin abnormalities of socio-emotional behaviour in major dementias. However, these processes remain poorly characterized. Here we used music to probe rule decoding and reward valuation in patients with frontotemporal dementia (FTD) syndromes and Alzheimer's disease (AD) relative to healthy age-matched individuals. We created short melodies that were...

Attention provides the gateway to cognition, by selecting certain stimuli for further analysis. Recent research demonstrates that whether a stimulus captures attention is not determined solely by its physical properties, but is malleable, being influenced by our previous experience of rewards obtained by attending to that stimulus. Here we show that this influence of reward learning on attentio...

Dopamine (DA) is believed to play a fundamental role in reward processes. Virtually all drugs of abuse activate dopaminergic systems, as do "natural" rewards such as sexual interaction and food. Sweet-tasting solutions, for example, are a well characterized natural reward. In the present experiments, we used mice that cannot make DA (DD mice) to test the hypothesis that DA is necessary for rewa...

Dopamine release in cortical and subcortical structures plays a central role in reward-related neural processes. Within this context, dopaminergic inputs are commonly assumed to play an activating role, facilitating behavioral and cognitive operations necessary to obtain a prospective reward. Here, we provide evidence from human fMRI that this activating role can also be mediated by task-demand...

We present a probabilistic algorithm for nonlinear inverse reinforcement learning. The goal of inverse reinforcement learning is to learn the reward function in a Markov decision process from expert demonstrations. While most prior inverse reinforcement learning algorithms represent the reward as a linear combination of a set of features, we use Gaussian processes to learn the reward as a nonli...