Reinforcement learning (RL) is a popular paradigm for addressing sequential decision tasks in which the agent has only limited environmental feedback. Despite many advances over the past three decades, learning in many domains still requires a large amount of interaction with the environment, which can be prohibitively expensive in realistic scenarios. To address this problem, transfer learning has been applied to reinforcement learning such that experience gained in one task can be leveraged when starting to learn the next, harder task. More recently, several lines of research have explored how tasks, or data samples themselves, can be sequenced into a curriculum for the purpose of learning a problem that may otherwise be too difficult to learn from scratch. In this article, we present a framework for curriculum learning (CL) in reinforcement learning, and use it to survey and classify existing CL methods in terms of their assumptions, capabilities, and goals. Finally, we use our framework to find open problems and suggest directions for future RL curriculum learning research.

Although deep learning models have achieved state-of-the art performance on a number of vision tasks, generalization over high dimensional multi-modal data, and reliable predictive uncertainty estimation are still active areas of research. Bayesian approaches including Bayesian Neural Nets (BNNs) do not scale well to modern computer vision tasks, as they are difficult to train, and have poor generalization under dataset-shift [27,38]. This motivates the need for effective ensembles which can generalize and give reliable uncertainty estimates. In this paper, we target the problem of generating effective ensembles of neural networks by encouraging diversity in prediction. We explicitly optimize a diversity inducing adversarial loss for learning the stochastic latent variables and thereby obtain diversity in the output predictions necessary for modeling multi-modal data. We evaluate our method on benchmark datasets: MNIST, CIFAR100, TinyImageNet and MIT Places 2, and compared to the most competitive baselines show significant improvements in classification accuracy, under a shift in the data distribution and in out-of-distribution detection.

Machine Learning is one of the most trending technologies available today. In this blog, you will learn about some of the most popular and widely used Machine Algorithms. However, let's first try to understand the meaning of Machine Learning and its algorithms. Machine Learning (ML) allows systems to gain knowledge from past information and experiences to improve their performance without being explicitly programmed. It uses Deep Learning and other advanced technologies in order to help the systems learn.

We consider causal discovery from time series using conditional independence (CI) based network learning algorithms such as the PC algorithm. The PC algorithm is divided into a skeleton phase where adjacencies are determined based on efficiently selected CI tests and subsequent phases where links are oriented utilizing the Markov and Faithfulness assumptions. Here we show that autocorrelation makes the PC algorithm much less reliable with very low adjacency and orientation detection rates and inflated false positives. We propose a new algorithm, called PCMCI$^+$ that extends the PCMCI method from [Runge et al., 2019b] to also include discovery of contemporaneous links. It separates the skeleton phase for lagged and contemporaneous conditioning sets and modifies the conditioning sets for the individual CI tests. We show that this algorithm now benefits from increasing autocorrelation and yields much more adjacency detection power and especially more orientation recall for contemporaneous links while controlling false positives and having much shorter runtimes. Numerical experiments indicate that the algorithm can be of considerable use in many application scenarios for dozens of variables and large time delays.

We investigate spatio-temporal event analysis using point processes. Inferring the dynamics of event sequences spatiotemporally has many practical applications including crime prediction, social media analysis, and traffic forecasting. In particular, we focus on spatio-temporal Hawkes processes that are commonly used due to their capability to capture excitations between event occurrences. We introduce a novel inference framework based on randomized transformations and gradient descent to learn the process. We replace the spatial kernel calculations by randomized Fourier feature-based transformations. The introduced randomization by this representation provides flexibility while modeling the spatial excitation between events. Moreover, the system described by the process is expressed within closed-form in terms of scalable matrix operations. During the optimization, we use maximum likelihood estimation approach and gradient descent while properly handling positivity and orthonormality constraints. The experiment results show the improvements achieved by the introduced method in terms of fitting capability in synthetic and real datasets with respect to the conventional inference methods in the spatio-temporal Hawkes process literature. We also analyze the triggering interactions between event types and how their dynamics change in space and time through the interpretation of learned parameters.

In this paper, as a study of reinforcement learning, we converge the Q function to unbounded rewards such as Gaussian distribution. From the central limit theorem, in some real-world applications it is natural to assume that rewards follow a Gaussian distribution , but existing proofs cannot guarantee convergence of the Q-function. Furthermore, in the distribution-type reinforcement learning and Bayesian reinforcement learning that have become popular in recent years, it is better to allow the reward to have a Gaussian distribution. Therefore, in this paper, we prove the convergence of the Q-function under the condition of $E[r(s,a)^2]<\infty$, which is much more relaxed than the existing research. Finally, as a bonus, a proof of the policy gradient theorem for distributed reinforcement learning is also posted.

Adversarial Machine Learning (AML) is emerging as a major field aimed at the protection of automated ML systems against security threats. The majority of work in this area has built upon a game-theoretic framework by modelling a conflict between an attacker and a defender. After reviewing game-theoretic approaches to AML, we discuss the benefits that a Bayesian Adversarial Risk Analysis perspective brings when defending ML based systems. A research agenda is included.

Approximate inference in probability models is a fundamental task in machine learning. Approximate inference provides powerful tools to Bayesian reasoning, decision making, and Bayesian deep learning. The main goal is to estimate the expectation of interested functions w.r.t. a target distribution. When it comes to high dimensional probability models and large datasets, efficient approximate inference becomes critically important. In this thesis, we propose a new framework for approximate inference, which combines the advantages of these three frameworks and overcomes their limitations. Our proposed four algorithms are motivated by the recent computational progress of Stein's method. Our proposed algorithms are applied to continuous and discrete distributions under the setting when the gradient information of the target distribution is available or unavailable. Theoretical analysis is provided to prove the convergence of our proposed algorithms. Our adaptive IS algorithm iteratively improves the importance proposal by functionally decreasing the KL divergence between the updated proposal and the target. When the gradient of the target is unavailable, our proposed sampling algorithm leverages the gradient of a surrogate model and corrects induced bias with importance weights, which significantly outperforms other gradient-free sampling algorithms. In addition, our theoretical results enable us to perform the goodness-of-fit test on discrete distributions. At the end of the thesis, we propose an importance-weighted method to efficiently aggregate local models in distributed learning with one-shot communication. Results on simulated and real datasets indicate the statistical efficiency and wide applicability of our algorithm.

Increasing the autonomy level of a robot hand to accomplish remote object manipulation tasks faster and easier is a new and promising topic in teleoperation. Such semi-autonomous telemanipulation, however, is very challenging due to the physical discrepancy between the human hand and the robot hand, along with the fine motion constraints required for the manipulation task. To overcome these challenges, the robot needs to learn how to assist the human operator in a preferred/intuitive way, which must provide effective assistance that the operator needs yet still accommodate human inputs, so the operator feels in control of the system (i.e., not counter-intuitive to the operator). Toward this goal, we develop novel data-driven approaches to stably learn what assistance is preferred from high data variance caused by the ambiguous nature of human operators. To avoid an extensive robot-specific training process, methods to transfer this assistance knowledge between different robot hands are discussed. Experiments were conducted to telemanipulate a cup for three principal tasks: usage, move, and handover by remotely controlling a 3-finger gripper and 2-finger gripper. Results demonstrated that the proposed model effectively learned the knowledge of preferred assistance, and knowledge transfer between robots allows this semi-autonomous telemanipulation strategy to be scaled up with less training efforts.

Deep energy-based models (EBMs) are very flexible in distribution parametrization but computationally challenging because of the intractable partition function. They are typically trained via maximum likelihood, using contrastive divergence to approximate the gradient of the KL divergence between data and model distribution. While KL divergence has many desirable properties, other f-divergences have shown advantages in training implicit density generative models such as generative adversarial networks. In this paper, we propose a general variational framework termed f-EBM to train EBMs using any desired f-divergence. We introduce a corresponding optimization algorithm and prove its local convergence property with non-linear dynamical systems theory. Experimental results demonstrate the superiority of f-EBM over contrastive divergence, as well as the benefits of training EBMs using f-divergences other than KL.