Efficient and effective exploration in continuous space is a central problem in applying reinforcement learning (RL) to autonomous driving. Skills learned from expert demonstrations or designed for specific tasks can benefit the exploration, but they are usually costly-collected, unbalanced/sub-optimal, or failing to transfer to diverse tasks. However, human drivers can adapt to varied driving tasks without demonstrations by taking efficient and structural explorations in the entire skill space rather than a limited space with task-specific skills. Inspired by the above fact, we propose an RL algorithm exploring all feasible motion skills instead of a limited set of task-specific and object-centric skills. Without demonstrations, our method can still perform well in diverse tasks. First, we build a task-agnostic and ego-centric (TaEc) motion skill library in a pure motion perspective, which is diverse enough to be reusable in different complex tasks. The motion skills are then encoded into a low-dimension latent skill space, in which RL can do exploration efficiently. Validations in various challenging driving scenarios demonstrate that our proposed method, TaEc-RL, outperforms its counterparts significantly in learning efficiency and task performance.

Recently, the seminal algorithms AlphaGo and AlphaZero have started a new era in game learning and deep reinforcement learning. While the achievements of AlphaGo and AlphaZero - playing Go and other complex games at super human level - are truly impressive, these architectures have the drawback that they require high computational resources. Many researchers are looking for methods that are similar to AlphaZero, but have lower computational demands and are thus more easily reproducible. In this paper, we pick an important element of AlphaZero - the Monte Carlo Tree Search (MCTS) planning stage - and combine it with temporal difference (TD) learning agents. We wrap MCTS for the first time around TD n-tuple networks and we use this wrapping only at test time to create versatile agents that keep at the same time the computational demands low. We apply this new architecture to several complex games (Othello, ConnectFour, Rubik's Cube) and show the advantages achieved with this AlphaZero-inspired MCTS wrapper. In particular, we present results that this agent is the first one trained on standard hardware (no GPU or TPU) to beat the very strong Othello program Edax up to and including level 7 (where most other learning-from-scratch algorithms could only defeat Edax up to level 2).

"Berkeley Lab is unique because its machine learning expertise is reasonably well established, and its tradition of team science means that we can work with researchers to apply these methods to scientific problems." "Although much of the time and effort spent in the software maintenance is not reflected in our research publication list, it is more than rewarding to see the wide use of this software in both the high-end scientific world and the commercial world." "I think one of the things Berkeley Lab does well is allow people to make collaborations that advance science much more efficiently." Berkeley Lab's research into machine learning builds on its foundational work in mathematics to develop methods that are consistent with physical laws, robust in the presence of noisy or biased data, and capable of being interpreted and explained in scientifically meaningful ways. Berkeley Lab Research Scientist Mariam Kiran uses deep reinforcement learning and innovative multi-objective optimization techniques to train network controllers to predict network traffic and improve traffic engineering.

Energy-function-based safety certificates can provide provable safety guarantees for the safe control tasks of complex robotic systems. However, all recent studies about learning-based energy function synthesis only consider the feasibility, which might cause over-conservativeness and result in less efficient controllers. In this work, we proposed the magnitude regularization technique to improve the efficiency of safe controllers by reducing the conservativeness inside the energy function while keeping the promising provable safety guarantees. Specifically, we quantify the conservativeness by the magnitude of the energy function, and we reduce the conservativeness by adding a magnitude regularization term to the synthesis loss. We propose the SafeMR algorithm that uses reinforcement learning (RL) for the synthesis to unify the learning processes of safe controllers and energy functions. Experimental results show that the proposed method does reduce the conservativeness of the energy functions and outperforms the baselines in terms of the controller efficiency while guaranteeing safety.

StarCraft II (SC2) poses a grand challenge for reinforcement learning (RL), of which the main difficulties include huge state space, varying action space, and a long time horizon. In this work, we investigate a set of RL techniques for the full-length game of StarCraft II. We investigate a hierarchical RL approach involving extracted macro-actions and a hierarchical architecture of neural networks. We investigate a curriculum transfer training procedure and train the agent on a single machine with 4 GPUs and 48 CPU threads. On a 64x64 map and using restrictive units, we achieve a win rate of 99% against the level-1 built-in AI. Through the curriculum transfer learning algorithm and a mixture of combat models, we achieve a 93% win rate against the most difficult non-cheating level built-in AI (level-7). In this extended version of the paper, we improve our architecture to train the agent against the cheating level AIs and achieve the win rate against the level-8, level-9, and level-10 AIs as 96%, 97%, and 94%, respectively. Our codes are at https://github.com/liuruoze/HierNet-SC2. To provide a baseline referring the AlphaStar for our work as well as the research and open-source community, we reproduce a scaled-down version of it, mini-AlphaStar (mAS). The latest version of mAS is 1.07, which can be trained on the raw action space which has 564 actions. It is designed to run training on a single common machine, by making the hyper-parameters adjustable. We then compare our work with mAS using the same resources and show that our method is more effective. The codes of mini-AlphaStar are at https://github.com/liuruoze/mini-AlphaStar. We hope our study could shed some light on the future research of efficient reinforcement learning on SC2 and other large-scale games.

Computers can understand and then engage with people in an emotionally intelligent way thanks to speech-emotion recognition (SER). However, the performance of SER in cross-corpus and real-world live data feed scenarios can be significantly improved. The inability to adapt an existing model to a new domain is one of the shortcomings of SER methods. To address this challenge, researchers have developed domain adaptation techniques that transfer knowledge learnt by a model across the domain. Although existing domain adaptation techniques have improved performances across domains, they can be improved to adapt to a real-world live data feed situation where a model can self-tune while deployed. In this paper, we present a deep reinforcement learning-based strategy (RL-DA) for adapting a pre-trained model to a real-world live data feed setting while interacting with the environment and collecting continual feedback. RL-DA is evaluated on SER tasks, including cross-corpus and cross-language domain adaption schema. Evaluation results show that in a live data feed setting, RL-DA outperforms a baseline strategy by 11% and 14% in cross-corpus and cross-language scenarios, respectively.

We propose learning via retracing, a novel self-supervised approach for learning the state representation (and the associated dynamics model) for reinforcement learning tasks. In addition to the predictive (reconstruction) supervision in the forward direction, we propose to include "retraced" transitions for representation / model learning, by enforcing the cycle-consistency constraint between the original and retraced states, hence improve upon the sample efficiency of learning. Moreover, learning via retracing explicitly propagates information about future transitions backward for inferring previous states, thus facilitates stronger representation learning for the downstream reinforcement learning tasks. We introduce Cycle-Consistency World Model (CCWM), a concrete model-based instantiation of learning via retracing. Additionally we propose a novel adaptive "truncation" mechanism for counteracting the negative impacts brought by "irreversible" transitions such that learning via retracing can be maximally effective. Through extensive empirical studies on visual-based continuous control benchmarks, we demonstrate that CCWM achieves state-of-the-art performance in terms of sample efficiency and asymptotic performance, whilst exhibiting behaviours that are indicative of stronger representation learning.

We consider a discounted cost constrained Markov decision process (CMDP) policy optimization problem, in which an agent seeks to maximize a discounted cumulative reward subject to a number of constraints on discounted cumulative utilities. To solve this constrained optimization program, we study an online actor-critic variant of a classic primal-dual method where the gradients of both the primal and dual functions are estimated using samples from a single trajectory generated by the underlying time-varying Markov processes. This online primal-dual natural actor-critic algorithm maintains and iteratively updates three variables: a dual variable (or Lagrangian multiplier), a primal variable (or actor), and a critic variable used to estimate the gradients of both primal and dual variables. These variables are updated simultaneously but on different time scales (using different step sizes) and they are all intertwined with each other. Our main contribution is to derive a finite-time analysis for the convergence of this algorithm to the global optimum of a CMDP problem. Specifically, we show that with a proper choice of step sizes the optimality gap and constraint violation converge to zero in expectation at a rate $\mathcal{O}(1/K^{1/6})$, where K is the number of iterations. To our knowledge, this paper is the first to study the finite-time complexity of an online primal-dual actor-critic method for solving a CMDP problem. We also validate the effectiveness of this algorithm through numerical simulations.

In recent years Landmark Complexes have been successfully employed for localization-free and metric-free autonomous exploration using a group of sensing-limited and communication-limited robots in a GPS-denied environment. To ensure rapid and complete exploration, existing works make assumptions on the density and distribution of landmarks in the environment. These assumptions may be overly restrictive, especially in hazardous environments where landmarks may be destroyed or completely missing. In this paper, we first propose a deep reinforcement learning framework for multi-agent cooperative exploration in environments with sparse landmarks while reducing client-server communication. By leveraging recent development on partial observability and credit assignment, our framework can train the exploration policy efficiently for multi-robot systems. The policy receives individual rewards from actions based on a proximity sensor with limited range and resolution, which is combined with group rewards to encourage collaborative exploration and construction of the Landmark Complex through observation of 0-, 1- and 2-dimensional simplices. In addition, we employ a three-stage curriculum learning strategy to mitigate the reward sparsity by gradually adding random obstacles and destroying random landmarks. Experiments in simulation demonstrate that our method outperforms the state-of-the-art landmark complex exploration method in efficiency among different environments with sparse landmarks.

Enabling the capability of assessing risk and making risk-aware decisions is essential to applying reinforcement learning to safety-critical robots like drones. In this paper, we investigate a specific case where a nano quadcopter robot learns to navigate an apriori-unknown cluttered environment under partial observability. We present a distributional reinforcement learning framework to generate adaptive risk-tendency policies. Specifically, we propose to use lower tail conditional variance of the learnt return distribution as intrinsic uncertainty estimation, and use exponentially weighted average forecasting (EWAF) to adapt the risk-tendency in accordance with the estimated uncertainty. In simulation and real-world empirical results, we show that (1) the most effective risk-tendency vary across states, (2) the agent with adaptive risk-tendency achieves superior performance compared to risk-neutral policy or risk-averse policy baselines.