Web servers scaled across distributed systems necessitate complex runtime controls for providing quality of service (QoS) guarantees as well as minimizing the energy costs under dynamic workloads. This paper presents a QoS-aware runtime controller using horizontal scaling (node allocation) and vertical scaling (resource allocation within nodes) methods synergistically to provide adaptation to workloads while minimizing the power consumption under QoS constraint (i.e., response time). A horizontal scaling determines the number of active nodes based on workload demands and the required QoS according to a set of rules. Then, it is coupled with vertical scaling using transfer Q-learning, which further tunes power/performance based on workload profile using dynamic voltage/frequency scaling (DVFS). It transfers Q-values within minimally explored states reducing exploration requirements. In addition, the approach exploits a scalable architecture of the many-core server allowing to reuse available knowledge from fully or partially explored nodes. When combined, these methods allow to reduce the exploration time and QoS violations when compared to model-free Q-learning. The technique balances design-time and runtime costs to maximize the portability and operational optimality demonstrated through persistent power reductions with minimal QoS violations under different workload scenarios on heterogeneous multi-processing nodes of a server cluster.

In many risk-aware and multi-objective reinforcement learning settings, the utility of the user is derived from the single execution of a policy. In these settings, making decisions based on the average future returns is not suitable. For example, in a medical setting a patient may only have one opportunity to treat their illness. When making a decision, just the expected return -- known in reinforcement learning as the value -- cannot account for the potential range of adverse or positive outcomes a decision may have. Our key insight is that we should use the distribution over expected future returns differently to represent the critical information that the agent requires at decision time. In this paper, we propose Distributional Monte Carlo Tree Search, an algorithm that learns a posterior distribution over the utility of the different possible returns attainable from individual policy executions, resulting in good policies for both risk-aware and multi-objective settings. Moreover, our algorithm outperforms the state-of-the-art in multi-objective reinforcement learning for the expected utility of the returns.

Sepsis is the leading cause of mortality in the ICU, responsible for 6% of all hospitalizations and 35% of all in-hospital deaths in USA. However, there is no universally agreed upon strategy for vasopressor and fluid administration. It has also been observed that different patients respond differently to treatment, highlighting the need for individualized treatment. Vasopressors and fluids are administrated with specific effects to cardiovascular physiology in mind and medical research has suggested that physiologic, hemodynamically guided, approaches to treatment. Thus we propose a novel approach, exploiting and unifying complementary strengths of Mathematical Modelling, Deep Learning, Reinforcement Learning and Uncertainty Quantification, to learn individualized, safe, and uncertainty aware treatment strategies. We first infer patient-specific, dynamic cardiovascular states using a novel physiology-driven recurrent neural network trained in an unsupervised manner. This information, along with a learned low dimensional representation of the patient's lab history and observable data, is then used to derive value distributions using Batch Distributional Reinforcement Learning. Moreover in a safety critical domain it is essential to know what our agent does and does not know, for this we also quantify the model uncertainty associated with each patient state and action, and propose a general framework for uncertainty aware, interpretable treatment policies. This framework can be tweaked easily, to reflect a clinician's own confidence of the framework, and can be easily modified to factor in human expert opinion, whenever it's accessible. Using representative patients and a validation cohort, we show that our method has learned physiologically interpretable generalizable policies.

Information theoretic sensor management approaches are an ideal solution to state estimation problems when considering the optimal control of multi-agent systems, however they are too computationally intensive for large state spaces, especially when considering the limited computational resources typical of large-scale distributed multi-agent systems. Reinforcement learning (RL) is a promising alternative which can find approximate solutions to distributed optimal control problems that take into account the resource constraints inherent in many systems of distributed agents. However, the RL training can be prohibitively inefficient, especially in low-information environments where agents receive little to no feedback in large portions of the state space. We propose a hybrid information-driven multi-agent reinforcement learning (MARL) approach that utilizes information theoretic models as heuristics to help the agents navigate large sparse state spaces, coupled with information based rewards in an RL framework to learn higher-level policies. This paper presents our ongoing work towards this objective. Our preliminary findings show that such an approach can result in a system of agents that are approximately three orders of magnitude more efficient at exploring a sparse state space than naive baseline metrics. While the work is still in its early stages, it provides a promising direction for future research.

In this paper, we propose a novel mechanism to normalize metamorphic and obfuscated malware down at the opcode level and hence create an advanced metamorphic malware de-obfuscation and defense system. We name this system DRLDO, for Deep Reinforcement Learning based De-Obfuscator. With the inclusion of the DRLDO as a sub-component, an existing Intrusion Detection System could be augmented with defensive capabilities against 'zero-day' attacks from obfuscated and metamorphic variants of existing malware. This gains importance, not only because there exists no system to date that uses advanced DRL to intelligently and automatically normalize obfuscation down even to the opcode level, but also because the DRLDO system does not mandate any changes to the existing IDS. The DRLDO system does not even mandate the IDS' classifier to be retrained with any new dataset containing obfuscated samples. Hence DRLDO could be easily retrofitted into any existing IDS deployment. We designed, developed, and conducted experiments on the system to evaluate the same against multiple-simultaneous attacks from obfuscations generated from malware samples from a standardized dataset that contains multiple generations of malware. Experimental results prove that DRLDO was able to successfully make the otherwise un-detectable obfuscated variants of the malware detectable by an existing pre-trained malware classifier. The detection probability was raised well above the cut-off mark to 0.6 for the classifier to detect the obfuscated malware unambiguously. Further, the de-obfuscated variants generated by DRLDO achieved a very high correlation (of 0.99) with the base malware. This observation validates that the DRLDO system is actually learning to de-obfuscate and not exploiting a trivial trick.

In this letter, we investigate the hybrid beamforming based on deep reinforcement learning (DRL) for millimeter Wave (mmWave) multi-user (MU) multiple-input-single-output (MISO) system. A multiagent DRL method is proposed to solve the exploration efficiency problem in DRL. In the proposed method, prioritized replay buffer and more informative reward are applied to accelerate the convergence. Simulation results show that the proposed architecture achieves higher spectral efficiency and less time consumption than the benchmarks, thus is more suitable for practical applications. To obtain the hybrid precoding matrices, several iterative methods, such as [1]-[4], have been proposed for single-user and multi-user (MU) systems.

To make good decisions in the real world people need efficient planning strategies because their computational resources are limited. Knowing which planning strategies would work best for people in different situations would be very useful for understanding and improving human decision-making. But our ability to compute those strategies used to be limited to very small and very simple planning tasks. To overcome this computational bottleneck, we introduce a cognitively-inspired reinforcement learning method that can overcome this limitation by exploiting the hierarchical structure of human behavior. The basic idea is to decompose sequential decision problems into two sub-problems: setting a goal and planning how to achieve it. This hierarchical decomposition enables us to discover optimal strategies for human planning in larger and more complex tasks than was previously possible. The discovered strategies outperform existing planning algorithms and achieve a super-human level of computational efficiency. We demonstrate that teaching people to use those strategies significantly improves their performance in sequential decision-making tasks that require planning up to eight steps ahead. By contrast, none of the previous approaches was able to improve human performance on these problems. These findings suggest that our cognitively-informed approach makes it possible to leverage reinforcement learning to improve human decision-making in complex sequential decision-problems. Future work can leverage our method to develop decision support systems that improve human decision making in the real world.

Endeavors for designing robots with human-level cognitive abilities have led to different categories of learning machines. According to Skinner's theory, reinforcement learning (RL) plays a key role in human intuition and cognition. Majority of the state-of-the-art methods including deep RL algorithms are strongly influenced by the connectionist viewpoint. Such algorithms can significantly benefit from theories of mind and learning in other disciplines. This paper entertains the idea that theories such as language of thought hypothesis (LOTH), script theory, and Piaget's cognitive development theory provide complementary approaches, which will enrich the RL field. Following this line of thinking, a general computational building block is proposed for Piaget's schema theory that supports the notions of productivity, systematicity, and inferential coherence as described by Fodor in contrast with the connectionism theory. Abstraction in the proposed method is completely upon the system itself and is not externally constrained by any predefined architecture. The whole process matches the Neisser's perceptual cycle model. Performed experiments on three typical control problems followed by behavioral analysis confirm the interpretability of the proposed method and its competitiveness compared to the state-of-the-art algorithms. Hence, the proposed framework can be viewed as a step towards achieving human-like cognition in artificial intelligent systems.

We study the regret of reinforcement learning from offline data generated by a fixed behavior policy in an infinite-horizon discounted Markov decision process (MDP). While existing analyses of common approaches, such as fitted $Q$-iteration (FQI), suggest a $O(1/\sqrt{n})$ convergence for regret, empirical behavior exhibits much faster convergence. In this paper, we present a finer regret analysis that exactly characterizes this phenomenon by providing fast rates for the regret convergence. First, we show that given any estimate for the optimal quality function $Q^*$, the regret of the policy it defines converges at a rate given by the exponentiation of the $Q^*$-estimate's pointwise convergence rate, thus speeding it up. The level of exponentiation depends on the level of noise in the decision-making problem, rather than the estimation problem. We establish such noise levels for linear and tabular MDPs as examples. Second, we provide new analyses of FQI and Bellman residual minimization to establish the correct pointwise convergence guarantees. As specific cases, our results imply $O(1/n)$ regret rates in linear cases and $\exp(-\Omega(n))$ regret rates in tabular cases.

This paper studies the control synthesis of motion planning subject to uncertainties. The uncertainties are considered in robot motion and environment properties, giving rise to the probabilistic labeled Markov decision process (MDP). A model-free reinforcement learning (RL) is developed to generate a finite-memory control policy to satisfy high-level tasks expressed in linear temporal logic (LTL) formulas. One of the novelties is to translate LTL into a limit deterministic generalized B\"uchi automaton (LDGBA) and develop a corresponding embedded LDGBA (E-LDGBA) by incorporating a tracking-frontier function to overcome the issue of sparse accepting rewards, resulting in improved learning performance without increasing computational complexity. Due to potentially conflicting tasks, a relaxed product MDP is developed to allow the agent to revise its motion plan without strictly following the desired LTL constraints if the desired tasks can only be partially fulfilled. An expected return composed of violation rewards and accepting rewards is developed. The designed violation function quantifies the differences between the revised and the desired motion planning, while the accepting rewards are designed to enforce the satisfaction of the acceptance condition of the relaxed product MDP. Rigorous analysis shows that any RL algorithm that optimizes the expected return is guaranteed to find policies that, in decreasing order, can 1) satisfy acceptance condition of relaxed product MDP and 2) reduce the violation cost over long-term behaviors. Also, we validate the control synthesis approach via simulation and experimental results.