A backdoor attack allows a malicious user to manipulate the environment or corrupt the training data, thus inserting a backdoor into the trained agent. Such attacks compromise the RL system's reliability, leading to potentially catastrophic results in various key fields. In contrast, relatively limited research has investigated effective defenses against backdoor attacks in RL. This paper proposes the Recovery Triggered States (RTS) method, a novel approach that effectively protects the victim agents from backdoor attacks. RTS involves building a surrogate network to approximate the dynamics model. Developers can then recover the environment from the triggered state to a clean state, thereby preventing attackers from activating backdoors hidden in the agent by presenting the trigger. When training the surrogate to predict states, we incorporate agent action information to reduce the discrepancy between the actions taken by the agent on predicted states and the actions taken on real states. RTS is the first approach to defend against backdoor attacks in a single-agent setting. Our results show that using RTS, the cumulative reward only decreased by 1.41% under the backdoor attack.

Many state-of-the art robotic applications utilize series elastic actuators (SEAs) with closed-loop force control to achieve complex tasks such as walking, lifting, and manipulation. Model-free PID control methods are more prone to instability due to nonlinearities in the SEA where cascaded model-based robust controllers can remove these effects to achieve stable force control. However, these model-based methods require detailed investigations to characterize the system accurately. Deep reinforcement learning (DRL) has proved to be an effective model-free method for continuous control tasks, where few works deal with hardware learning. This paper describes the training process of a DRL policy on hardware of an SEA pendulum system for tracking force control trajectories from 0.05 - 0.35 Hz at 50 N amplitude using the Proximal Policy Optimization (PPO) algorithm. Safety mechanisms are developed and utilized for training the policy for 12 hours (overnight) without an operator present within the full 21 hours training period. The tracking performance is evaluated showing improvements of $25$ N in mean absolute error when comparing the first 18 min. of training to the full 21 hours for a 50 N amplitude, 0.1 Hz sinusoid desired force trajectory. Finally, the DRL policy exhibits better tracking and stability margins when compared to a model-free PID controller for a 50 N chirp force trajectory.

Robotic caregivers could potentially improve the quality of life of many who require physical assistance. However, in order to assist individuals who are lying in bed, robots must be capable of dealing with a significant obstacle: the blanket or sheet that will almost always cover the person's body. We propose a method for targeted bedding manipulation over people lying supine in bed where we first learn a model of the cloth's dynamics. Then, we optimize over this model to uncover a given target limb using information about human body shape and pose that only needs to be provided at run-time. We show how this approach enables greater robustness to variation relative to geometric and reinforcement learning baselines via a number of generalization evaluations in simulation and in the real world. We further evaluate our approach in a human study with 12 participants where we demonstrate that a mobile manipulator can adapt to real variation in human body shape, size, pose, and blanket configuration to uncover target body parts without exposing the rest of the body. Source code and supplementary materials are available online.

Research Internships at Microsoft provide a dynamic environment for research careers with a network of world-class research labs led by globally-recognized scientists and engineers. Our researchers and engineers pursue innovation in a range of scientific and technical disciplines to help solve complex challenges in diverse fields, including computing, healthcare, economics, and the environment. The Machine Learning and Statistics and AutoML groups at Microsoft Research New England are hiring multiple interns to work alongside leading researchers and engineers to advance the state-of-the-art in the field. Projects include applied and theoretical machine learning research on topics including (but not limited to): auto-ML, transfer learning, domain adaptation, program synthesis, meta-learning, sign language modeling, statistics, approximate inference, distribution compression, weather and climate forecasting, causal inference, causal machine learning, ML for health, adversarial ML, learning and incentives, reinforcement learning, multi-agent reinforcement learning, deep learning, supervised learning, and unsupervised learning. Responsibilities: Interns put inquiry and theory into practice.

Existing approaches for autonomous control of pan-tilt-zoom (PTZ) cameras use multiple stages where object detection and localization are performed separately from the control of the PTZ mechanisms. These approaches require manual labels and suffer from performance bottlenecks due to error propagation across the multi-stage flow of information. The large size of object detection neural networks also makes prior solutions infeasible for real-time deployment in resource-constrained devices. We present an end-to-end deep reinforcement learning (RL) solution called Eagle to train a neural network policy that directly takes images as input to control the PTZ camera. Training reinforcement learning is cumbersome in the real world due to labeling effort, runtime environment stochasticity, and fragile experimental setups. We introduce a photo-realistic simulation framework for training and evaluation of PTZ camera control policies. Eagle achieves superior camera control performance by maintaining the object of interest close to the center of captured images at high resolution and has up to 17% more tracking duration than the state-of-the-art. Eagle policies are lightweight (90x fewer parameters than Yolo5s) and can run on embedded camera platforms such as Raspberry PI (33 FPS) and Jetson Nano (38 FPS), facilitating real-time PTZ tracking for resource-constrained environments. With domain randomization, Eagle policies trained in our simulator can be transferred directly to real-world scenarios.

Temporal knowledge graphs (TKGs) model the temporal evolution of events and have recently attracted increasing attention. Since TKGs are intrinsically incomplete, it is necessary to reason out missing elements. Although existing TKG reasoning methods have the ability to predict missing future events, they fail to generate explicit reasoning paths and lack explainability. As reinforcement learning (RL) for multi-hop reasoning on traditional knowledge graphs starts showing superior explainability and performance in recent advances, it has opened up opportunities for exploring RL techniques on TKG reasoning. However, the performance of RL-based TKG reasoning methods is limited due to: (1) lack of ability to capture temporal evolution and semantic dependence jointly; (2) excessive reliance on manually designed rewards. To overcome these challenges, we propose an adaptive reinforcement learning model based on attention mechanism (DREAM) to predict missing elements in the future. Specifically, the model contains two components: (1) a multi-faceted attention representation learning method that captures semantic dependence and temporal evolution jointly; (2) an adaptive RL framework that conducts multi-hop reasoning by adaptively learning the reward functions. Experimental results demonstrate DREAM outperforms state-of-the-art models on public dataset

An efficient team is essential for the company to successfully complete new projects. To solve the team formation problem considering person-job matching (TFP-PJM), a 0-1 integer programming model is constructed, which considers both person-job matching and team members' willingness to communicate on team efficiency, with the person-job matching score calculated using intuitionistic fuzzy numbers. Then, a reinforcement learning-assisted genetic programming algorithm (RL-GP) is proposed to enhance the quality of solutions. The RL-GP adopts the ensemble population strategies. Before the population evolution at each generation, the agent selects one from four population search modes according to the information obtained, thus realizing a sound balance of exploration and exploitation. In addition, surrogate models are used in the algorithm to evaluate the formation plans generated by individuals, which speeds up the algorithm learning process. Afterward, a series of comparison experiments are conducted to verify the overall performance of RL-GP and the effectiveness of the improved strategies within the algorithm. The hyper-heuristic rules obtained through efficient learning can be utilized as decision-making aids when forming project teams. This study reveals the advantages of reinforcement learning methods, ensemble strategies, and the surrogate model applied to the GP framework. The diversity and intelligent selection of search patterns along with fast adaptation evaluation, are distinct features that enable RL-GP to be deployed in real-world enterprise environments.

The graph colouring problem consists of assigning labels, or colours, to the vertices of a graph such that no two adjacent vertices share the same colour. In this work we investigate whether deep reinforcement learning can be used to discover a competitive construction heuristic for graph colouring. Our proposed approach, ReLCol, uses deep Q-learning together with a graph neural network for feature extraction, and employs a novel way of parameterising the graph that results in improved performance. Using standard benchmark graphs with varied topologies, we empirically evaluate the benefits and limitations of the heuristic learned by ReLCol relative to existing construction algorithms, and demonstrate that reinforcement learning is a promising direction for further research on the graph colouring problem.

We extend the provably convergent Full Gradient DQN algorithm for discounted reward Markov decision processes from Avrachenkov et al. (2021) to average reward problems. We experimentally compare widely used RVI Q-Learning with recently proposed Differential Q-Learning in the neural function approximation setting with Full Gradient DQN and DQN. We also extend this to learn Whittle indices for Markovian restless multi-armed bandits. We observe a better convergence rate of the proposed Full Gradient variant across different tasks.

Join the Discord Server to collaborate. This is a Deep Reinforcement Learning library focused on improving development by introducing RLOps - MLOps for reinforcement learning. This library is initially focused on reducing the time taken for training models and hyperparameter optimization (HPO) by pioneering evolutionary HPO techniques for reinforcement learning. Evolutionary HPO has been shown to drastically reduce overall training times by automatically converging on optimal hyperparameters, without requiring numerous training runs. We are constantly adding more algorithms, with a view to add hierarchical and multi-agent algorithms soon.