In real-world multi-robot systems, performing high-quality, collaborative behaviors requires robots to asynchronously reason about high-level action selection at varying time durations. Macro-Action Decentralized Partially Observable Markov Decision Processes (MacDec-POMDPs) provide a general framework for asynchronous decision making under uncertainty in fully cooperative multi-agent tasks. However, multi-agent deep reinforcement learning methods have only been developed for (synchronous) primitive-action problems. This paper proposes two Deep Q-Network (DQN) based methods for learning decentralized and centralized macro-action-value functions with novel macro-action trajectory replay buffers introduced for each case. Evaluations on benchmark problems and a larger domain demonstrate the advantage of learning with macro-actions over primitive-actions and the scalability of our approaches.

We consider three modern roles for logic in artificial intelligence, which are based on the theory of tractable Boolean circuits: (1) logic as a basis for computation, (2) logic for learning from a combination of data and knowledge, and (3) logic for reasoning about the behavior of machine learning systems.

Over this decade, an increasing number of processes is being automated through classification algorithms, being essential that these are robust and reliable if we are to trust key operations based on their output. State-of-the-art classifiers perform extraordinarily well on standard data, but they have been shown to be vulnerable to adversarial examples, data instances specifically targeted at fooling the algorithms (Comiter, 2019). As a fundamental hypothesis, algorithms rely on the use of independent and identically distributed (iid) data for both the training and test phases. However, security aspects in classification, which form part of the field of adversarial machine learning (AML), question such hypothesis due to the presence of adversaries ready to modify the data to obtain a benefit and, thus, making both distributions differ. Stemming from the pioneering work in adversarial classification (AC) in Dalvi et al. (2004), the paradigm used to model the confrontation between adversaries and classification systems has been game theory, see recent reviews in Biggio and Roli (2018) and Zhou et al. (2018). As an example, the most popular attacks, including the fast gradient sign method (FGSM) (Goodfellow et al., 2014b), may be viewed from a game-theoretic perspective. Similarly, two of the most promising defence techniques, adversarial training (AT) (Madry et al., 2018), which trains the defender model with attacked samples, and adversarial logit pairing (ALP) (Kannan et al., 2018), which encourages the logits of the model to be the same for both standard and adversarial inputs, may be framed in game theoretic terms. This perspective typically entails common knowledge hypothesis (Hargreaves-Heap and Varoufakis, 2004) which, from a fundamental point of view, are not sustainable in settings such as security, as adversaries try to hide and conceal information. Recent work (Naveiro et al., 2019) presented ACRA, a novel approach for AC based on Adversarial Risk

In this paper, we formulate the adaptive learning problem---the problem of how to find an individualized learning plan (called policy) that chooses the most appropriate learning materials based on learner's latent traits---faced in adaptive learning systems as a Markov decision process (MDP). We assume latent traits to be continuous with an unknown transition model. We apply a model-free deep reinforcement learning algorithm---the deep Q-learning algorithm---that can effectively find the optimal learning policy from data on learners' learning process without knowing the actual transition model of the learners' continuous latent traits. To efficiently utilize available data, we also develop a transition model estimator that emulates the learner's learning process using neural networks. The transition model estimator can be used in the deep Q-learning algorithm so that it can more efficiently discover the optimal learning policy for a learner. Numerical simulation studies verify that the proposed algorithm is very efficient in finding a good learning policy, especially with the aid of a transition model estimator, it can find the optimal learning policy after training using a small number of learners.

Traditional centralized multi-agent reinforcement learning (MARL) algorithms are sometimes unpractical in complicated applications, due to non-interactivity between agents, curse of dimensionality and computation complexity. Hence, several decentralized MARL algorithms are motivated. However, existing decentralized methods only handle the fully cooperative setting where massive information needs to be transmitted in training. The block coordinate gradient descent scheme they used for successive independent actor and critic steps can simplify the calculation, but it causes serious bias. In this paper, we propose a flexible fully decentralized actor-critic MARL framework, which can combine most of actor-critic methods, and handle large-scale general cooperative multi-agent setting. A primal-dual hybrid gradient descent type algorithm framework is designed to learn individual agents separately for decentralization. From the perspective of each agent, policy improvement and value evaluation are jointly optimized, which can stabilize multi-agent policy learning. Furthermore, our framework can achieve scalability and stability for large-scale environment and reduce information transmission, by the parameter sharing mechanism and a novel modeling-other-agents methods based on theory-of-mind and online supervised learning. Sufficient experiments in cooperative Multi-agent Particle Environment and StarCraft II show that our decentralized MARL instantiation algorithms perform competitively against conventional centralized and decentralized methods.

Finite mixture models are an important tool in the statistical analysis of data, for example in data clustering. The optimal parameters of a mixture model are usually computed by maximizing the log-likelihood functional via the Expectation-Maximization algorithm. We propose an alternative approach based on the theory of Mean Field Games, a class of differential games with an infinite number of agents. We show that the solution of a finite state space multi-population Mean Field Games system characterizes the critical points of the log-likelihood functional for a Bernoulli mixture. The approach is then generalized to mixture models of categorical distributions. Hence, the Mean Field Games approach provides a method to compute the parameters of the mixture model, and we show its application to some standard examples in cluster analysis.

Biomedical data is filled with continuous real values; these values in the feature set tend to create problems like underfitting, the curse of dimensionality and increase in misclassification rate because of higher variance. In response, pre-processing techniques on dataset minimizes the side effects and have shown success in maintaining the adequate accuracy. Feature selection and discretization are the two necessary preprocessing steps that were effectively employed to handle the data redundancies in the biomedical data. However, in the previous works, the absence of unified effort by integrating feature selection and discretization together in solving the data redundancy problem leads to the disjoint and fragmented field. This paper proposes a novel multi-objective based dimensionality reduction framework, which incorporates both discretization and feature reduction as an ensemble model for performing feature selection and discretization. Selection of optimal features and the categorization of discretized and non-discretized features from the feature subset is governed by the multi-objective genetic algorithm (NSGA-II). The two objective, minimizing the error rate during the feature selection and maximizing the information gain while discretization is considered as fitness criteria.

Two-thirds of the people who buy a new car prefer to use a substitute instead of the built-in navigation system. However, for many applications, knowledge about a user's intended destination and route is crucial. For example, suggestions for available parking spots close to the destination can be made or ride-sharing opportunities along the route are facilitated. Our approach predicts probable destinations and routes of a vehicle, based on the most recent partial trajectory and additional contextual data. The approach follows a three-step procedure: First, a $k$-d tree-based space discretization is performed, mapping GPS locations to discrete regions. Secondly, a recurrent neural network is trained to predict the destination based on partial sequences of trajectories. The neural network produces destination scores, signifying the probability of each region being the destination. Finally, the routes to the most probable destinations are calculated. To evaluate the method, we compare multiple neural architectures and present the experimental results of the destination prediction. The experiments are based on two public datasets of non-personalized, timestamped GPS locations of taxi trips. The best performing models were able to predict the destination of a vehicle with a mean error of 1.3 km and 1.43 km respectively.

Reinforcement learning (RL) has been demonstrated to have great potential in many applications of scientific discovery and design. Recent work includes, for example, the design of new structures and compositions of molecules for therapeutic drugs. Much of the existing work related to the application of RL to scientific domains, however, assumes that the available state representation obeys the Markov property. For reasons associated with time, cost, sensor accuracy, and gaps in scientific knowledge, many scientific design and discovery problems do not satisfy the Markov property. Thus, something other than a Markov decision process (MDP) should be used to plan / find the optimal policy. In this paper, we present a physics-inspired semi-Markov RL environment, namely the phase change environment. In addition, we evaluate the performance of value-based RL algorithms for both MDPs and partially observable MDPs (POMDPs) on the proposed environment. Our results demonstrate deep recurrent Q-networks (DRQN) significantly outperform deep Q-networks (DQN), and that DRQNs benefit from training with hindsight experience replay. Implications for the use of semi-Markovian RL and POMDPs for scientific laboratories are also discussed.

Equity research analysts at financial institutions play a pivotal role in capital markets; they provide an efficient conduit between investors and companies' management and facilitate the efficient flow of information from companies, promoting functional and liquid markets. However, previous research in the academic finance and behavioral economics communities has found that analysts' estimates of future company earnings and other financial quantities can be affected by a number of behavioral, incentive-based and discriminatory biases and systematic errors, which can detrimentally affect both investors and public companies. We propose a Bayesian latent variable model for analysts' systematic errors and biases which we use to generate a robust bias-adjusted consensus estimate of company earnings. Experiments using historical earnings estimates data show that our model is more accurate than the consensus average of estimates and other related approaches.