Within 10 minutes of its birth, a baby fawn is able to stand. Within seven hours, it is able to walk. Between those two milestones, it engages in a highly adorable, highly frenetic flailing of limbs to figure it all out.

The intuitive collaboration of humans and intelligent robots (embodied AI) in the real-world is an essential objective for many desirable applications of robotics. Whilst there is much research regarding explicit communication, we focus on how humans and robots interact implicitly, on motor adaptation level. We present a real-world setup of a human-robot collaborative maze game, designed to be non-trivial and only solvable through collaboration, by limiting the actions to rotations of two orthogonal axes, and assigning each axes to one player. This results in neither the human nor the agent being able to solve the game on their own. We use a state-of-the-art reinforcement learning algorithm for the robotic agent, and achieve results within 30 minutes of real-world play, without any type of pre-training. We then use this system to perform systematic experiments on human/agent behaviour and adaptation when co-learning a policy for the collaborative game. We present results on how co-policy learning occurs over time between the human and the robotic agent resulting in each participant's agent serving as a representation of how they would play the game. This allows us to relate a person's success when playing with different agents than their own, by comparing the policy of the agent with that of their own agent.

Reinforcement learning (RL) algorithms often require expensive manual or automated hyperparameter searches in order to perform well on a new domain. This need is particularly acute in modern deep RL architectures which often incorporate many modules and multiple loss functions. In this paper, we take a step towards addressing this issue by using metagradients (Xu et al., 2018) to tune these hyperparameters via differentiable cross validation, whilst the agent interacts with and learns from the environment. We present the Self-Tuning Actor Critic (STAC) which uses this process to tune the hyperparameters of the usual loss function of the IMPALA actor critic agent(Espeholt et. al., 2018), to learn the hyperparameters that define auxiliary loss functions, and to balance trade offs in off policy learning by introducing and adapting the hyperparameters of a novel leaky V-trace operator. The method is simple to use, sample efficient and does not require significant increase in compute. Ablative studies show that the overall performance of STAC improves as we adapt more hyperparameters. When applied to 57 games on the Atari 2600 environment over 200 million frames our algorithm improves the median human normalized score of the baseline from 243% to 364%.

We propose a novel actor-critic, model-free reinforcement learning algorithm which employs a Bayesian method of parameter space exploration to solve environments. A Gaussian process is used to learn the expected return of a policy given the policy's parameters. The system is trained by updating the parameters using gradient descent on a new surrogate loss function consisting of the Proximal Policy Optimization 'Clipped' loss function and a bonus term representing the expected improvement acquisition function given by the Gaussian process. This new method is shown to be comparable to and at times empirically outperform current algorithms on environments that simulate robotic locomotion using the MuJoCo physics engine.

In this paper, we present a new paradigm to zero-shot learning (ZSL) that is trained by utilizing additional information (such as attribute-class mapping) for specific set of unseen classes. We conjecture that such additional information about unseen classes is more readily available than unsupervised image sets. Further, on close examination of the underlying attribute predictors of popular ZSL algorithms, we find that they often leverage attribute correlations to make predictions. While attribute correlations that remain intact in the unseen classes (test) benefit the prediction of difficult attributes, change in correlations can have an adverse effect on ZSL performance. For example, detecting an attribute 'brown' may be the same as detecting 'fur' over an animals' image dataset captured in the tropics. However, such a model might fail on unseen images of Arctic animals. To address this effect, termed target-shift in ZSL, we utilize our proposed framework to design grouped adversarial learning. We introduce grouping of attributes to enable the model to continue to benefit from useful correlations, while restricting cross-group correlations that may be harmful for generalization. Our analysis shows that it is possible to not only constrain the model from leveraging unwanted correlations, but also adjust them to specific test setting using only the additional information (the already available attribute-class mapping). We show empirical results for zero-shot predictions on standard benchmark datasets, namely, aPY, AwA2, SUN and CUB datasets. We further introduce to the research community, a new experimental train-test split that maximizes target-shift to further study its effects.

We consider online learning for episodic Markov decision processes (MDPs) with stochastic long-term budget constraints, which plays a central role in ensuring the safety of reinforcement learning. Here the loss function can vary arbitrarily across the episodes, whereas both the loss received and the budget consumption are revealed at the end of each episode. Previous works solve this problem under the restrictive assumption that the transition model of the MDP is known a priori and establish regret bounds that depend polynomially on the cardinalities of the state space $\mathcal{S}$ and the action space $\mathcal{A}$. In this work, we propose a new \emph{upper confidence primal-dual} algorithm, which only requires the trajectories sampled from the transition model. In particular, we prove that the proposed algorithm achieves $\tilde{\mathcal{O}}(L|\mathcal{S}|\sqrt{|\mathcal{A}|T})$ upper bounds of both the regret and the constraint violation, where $L$ is the length of each episode. Our analysis incorporates a new high-probability drift analysis of Lagrange multiplier processes into the celebrated regret analysis of upper confidence reinforcement learning, which demonstrates the power of "optimism in the face of uncertainty" in constrained online learning.

Reinforcement learning (RL) is attracting increasing interests in autonomous driving due to its potential to solve complex classification and control problems. However, existing RL algorithms are rarely applied to real vehicles for two predominant problems: behaviours are unexplainable, and they cannot guarantee safety under new scenarios. This paper presents a safe RL algorithm, called Parallel Constrained Policy Optimization (PCPO), for two autonomous driving tasks. PCPO extends today's common actor-critic architecture to a three-component learning framework, in which three neural networks are used to approximate the policy function, value function and a newly added risk function, respectively. Meanwhile, a trust region constraint is added to allow large update steps without breaking the monotonic improvement condition. To ensure the feasibility of safety constrained problems, synchronized parallel learners are employed to explore different state spaces, which accelerates learning and policy-update. The simulations of two scenarios for autonomous vehicles confirm we can ensure safety while achieving fast learning.

Emergency vehicle (EV) service is a key function of cities and is exceedingly challenging due to urban traffic congestion. A key contributor to EV service delay is the lack of communication and cooperation between vehicles blocking EVs. In this paper, we study the improvement of EV service using vehicle-to-vehicle connectivity. We consider the establishment of dynamic queue jumper lanes (DQJLs) based on real-time coordination of connected vehicles. We develop a novel stochastic dynamic programming formulation for the DQJL problem, which explicitly account for the uncertainty of drivers' reaction to approaching EVs. We propose a deep neural network-based approximate dynamic programming (ADP) algorithm that efficiently computes the optimal coordination instructions. We also validate our approach on a micro-simulation testbed using Simulation On Urban Mobility (SUMO).

Regular Decision Processes (RDPs) are a recently introduced model that extends MDPs with non-Markovian dynamics and rewards. The non-Markovian behavior is restricted to depend on regular properties of the history. These can be specified using regular expressions or formulas in linear dynamic logic over finite traces. Fully specified RDPs can be solved by compiling them into an appropriate MDP. Learning RDPs from data is a challenging problem that has yet to be addressed, on which we focus in this paper. Our approach rests on a new representation for RDPs using Mealy Machines that emit a distribution and an expected reward for each state-action pair. Building on this representation, we combine automata learning techniques with history clustering to learn such a Mealy machine and solve it by adapting MCTS to it. We empirically evaluate this approach, demonstrating its feasibility.

Tuning hyper-parameters for evolutionary algorithms is an important issue in computational intelligence. Performance of an evolutionary algorithm depends not only on its operation strategy design, but also on its hyper-parameters. Hyper-parameters can be categorized in two dimensions as structural/numerical and time-invariant/time-variant. Particularly, structural hyper-parameters in existing studies are usually tuned in advance for time-invariant parameters, or with hand-crafted scheduling for time-invariant parameters. In this paper, we make the first attempt to model the tuning of structural hyper-parameters as a reinforcement learning problem, and present to tune the structural hyper-parameter which controls computational resource allocation in the CEC 2018 winner algorithm by Q-learning. Experimental results show favorably against the winner algorithm on the CEC 2018 test functions.