Accurate and reliable prediction of hospital admission location is important due to resource-constraints and space availability in a clinical setting, particularly when dealing with patients who come from the emergency department. In this work we propose a student-teacher network via reinforcement learning to deal with this specific problem. A representation of the weights of the student network is treated as the state and is fed as an input to the teacher network. The teacher network's action is to select the most appropriate batch of data to train the student network on from a training set sorted according to entropy. By validating on three datasets, not only do we show that our approach outperforms state-of-the-art methods on tabular data and performs competitively on image recognition, but also that novel curricula are learned by the teacher network. We demonstrate experimentally that the teacher network can actively learn about the student network and guide it to achieve better performance than if trained alone.

Reinforcement learning studies sequential decision making problems where a learning agent repeatedly interacts with an environment and aims to improve her strategy over time based on the received feedback. One of the most fundamental tradeoffs in reinforcement learning theory is the exploration vs. exploitation tradeoff, that arises whenever the learner observes only partial feedback after each of her decisions, thus having to balance between exploring new strategies and exploiting those that are already known to perform well. The most basic and well-studied form of partial feedback is the so-called "bandit" feedback, where the learner only observes the cost of her chosen action on each decision round, while obtaining no information about the performance of other actions. Traditionally, the environment dynamics in reinforcement learning are modeled as a Markov Decision Process (MDP) with a finite number of possible states and actions. The MDP model has been studied and analyzed in numerous different settings and under various assumptions on the transition parameters, the nature of the reward functions, and the feedback model. Recently, a particular focus has been given to continuous state-action MDPs, and in particular, to a specific family of models in classic control where the state transition function is linear.

We are interested in how to design reinforcement learning agents that provably reduce the sample complexity for learning new tasks by transferring knowledge from previously-solved ones. The availability of solutions to related problems poses a fundamental trade-off: whether to seek policies that are expected to achieve high (yet sub-optimal) performance in the new task immediately or whether to seek information to quickly identify an optimal solution, potentially at the cost of poor initial behavior. In this work, we focus on the second objective when the agent has access to a generative model of state-action pairs. First, given a set of solved tasks containing an approximation of the target one, we design an algorithm that quickly identifies an accurate solution by seeking the state-action pairs that are most informative for this purpose. We derive PAC bounds on its sample complexity which clearly demonstrate the benefits of using this kind of prior knowledge. Then, we show how to learn these approximate tasks sequentially by reducing our transfer setting to a hidden Markov model and employing spectral methods to recover its parameters. Finally, we empirically verify our theoretical findings in simple simulated domains.

Learning from Demonstrations (LfD) is an active research area that addresses the problem of learning how to perform a task by observing the demonstrations provided by an expert. This approach plays an important role in many real-life learning settings, including human-to-robot interaction [1, 2, 3, 4, 5]. The two popular approaches for LfD include (i) behavioral cloning, which directly mimics the expert behavior, without understanding the objective [6], and (ii) inverse reinforcement learning (IRL), which infers the reward function (i.e., the objective of the task) explaining the expert behavior [7]. In this work, we focus on the IRL approach to LfD. Typically, the IRL learner assumes that the demonstrated expert behavior is optimal with respect to some reward function, even if the reward function cannot be specified explicitly as in typical reinforcement learning (RL).

Robustness of machine learning methods is essential for modern practical applications. Given the arms race between attack and defense methods, one may be curious regarding the fundamental limits of any defense mechanism. In this work, we focus on the problem of learning from noise-injected data, where the existing literature falls short by either assuming a specific attack method or by over-specifying the learning problem. We shed light on the information-theoretic limits of adversarial learning without assuming a particular learning process or attacker. Finally, we apply our general bounds to a canonical set of non-trivial learning problems and provide examples of common types of attacks.

Personalized marketing for retail consumers and account-based marketing for B2B customers now have proven value. Online interactions with customers generate large volumes of data for granular learning about consumer behavior for customization of product recommendations, messages, and content. The missing piece is a scalable and just-in-time way to gauge customer preferences and make product recommendations while visitors engage with websites. Deep reinforcement learning algorithms have been trained at the threshold level where they begin to achieve conversion rates to match the costs of data analysis. The touchstone of reinforcement learning (RL) is that it experiments with multiple pathways to achieve the objective of acquiring customers or any other goal.

Robots have been useful in environments that can be carefully controlled, such as those commonly found in industrial settings (e.g. assembly lines). However, in unstructured settings like the home, we need robotic systems that are adaptive to the diversity of the real world. Learning-based algorithms have the potential to enable robots to acquire complex behaviors adaptively in unstructured environments, by leveraging data collected from the environment. In particular, with reinforcement learning, robots learn novel behaviors through trial and error interactions. This is particularly important as we deploy robots in scenarios where the environment may not be known.

The brain of a human child is spectacularly amazing. Even in any previously unknown situation, the brain makes a decision based on its primal knowledge. Depending on the outcome, it learns and remembers the most optimal choices to be taken in that particular scenario. On a high level, this process of learning can be understood as a ’trial and error’ process, where the brain tries to maximise the occurrence of positive outcomes.

In a world saturated by artificial intelligence, Machine Learning, and over-zealous talks about both, it is important to understand and identify the types of Machine Learning we may encounter. For the practitioners creating these algorithms, it's essential to know the types of machine learning so that for any given task they may encounter, they can craft the proper learning environment and understand what to apply. Machine Learning can be broadly classified into 3 types:- Supervised learning, Unsupervised learning and Reinforcement Learning. Supervised learning is the most popular paradigm for machine learning. It is the easiest to understand and the simplest to implement.

Deep Deterministic Policy Gradient (DDPG) algorithm is one of the most well-known reinforcement learning methods. However, this method is inefficient and unstable in practical applications. On the other hand, the bias and variance of the Q estimation in the target function are sometimes difficult to control. This paper proposes a Regularly Updated Deterministic (RUD) policy gradient algorithm for these problems. This paper theoretically proves that the learning procedure with RUD can make better use of new data in replay buffer than the traditional procedure. In addition, the low variance of the Q value in RUD is more suitable for the current Clipped Double Q-learning strategy. This paper has designed a comparison experiment against previous methods, an ablation experiment with the original DDPG, and other analytical experiments in Mujoco environments. The experimental results demonstrate the effectiveness and superiority of RUD.