Among approaches for provably safe reinforcement learning, Model Predictive Shielding (MPS) has proven effective at complex tasks in continuous, high-dimensional state spaces, by leveraging a to ensure safety when the learned policy attempts to take risky actions. However, while MPS can ensure safety both during and after training, it often hinders task progress due to the conservative and task-oblivious nature of backup policies.This paper introduces (DMPS), which optimizes reinforcement learning objectives while maintaining provable safety. DMPS employs a local planner to dynamically select safe recovery actions that maximize both short-term progress as well as long-term rewards. Crucially, the planner and the neural policy play a synergistic role in DMPS. When planning recovery actions for ensuring safety, the planner utilizes the neural policy to estimate long-term rewards, allowing it to beyond its short-term planning horizon. Conversely, the neural policy under training learns from the recovery plans proposed by the planner, converging to policies that are both and in practice.This approach guarantees safety during and after training, with bounded recovery regret that decreases exponentially with planning horizon depth. Experimental results demonstrate that DMPS converges to policies that rarely require shield interventions after training and achieve higher rewards compared to several state-of-the-art baselines.

Recent advances in robotic manipulation have integrated low-level robotic control into Vision-Language Models (VLMs), extending them into Vision-Language-Action (VLA) models. Although state-of-the-art VLAs achieve strong performance in downstream robotic applications, supported by large-scale crowd-sourced robot training data, they still inevitably encounter failures during execution. Enabling robots to reason and recover from unpredictable and abrupt failures remains a critical challenge. Existing robotic manipulation datasets, collected in either simulation or the real world, primarily provide only ground-truth trajectories, leaving robots unable to recover once failures occur. Moreover, the few datasets that address failure detection typically offer only textual explanations, which are difficult to utilize directly in VLA models. To address this gap, we introduce FailSafe, a novel failure generation and recovery system that automatically produces diverse failure cases paired with executable recovery actions. FailSafe can be seamlessly applied to any manipulation task in any simulator, enabling scalable creation of failure action data. To demonstrate its effectiveness, we fine-tune LLaVa-OneVision-7B (LLaVa-OV-7B) to build FailSafe-VLM. Experimental results show that FailSafe-VLM successfully helps robotic arms detect and recover from potential failures, improving the performance of three state-of-the-art VLA models (pi0-FAST, OpenVLA, OpenVLA-OFT) by up to 22.6% on average across several tasks in Maniskill. Furthermore, FailSafe-VLM could generalize across different spatial configurations, camera viewpoints, object and robotic embodiments. We plan to release the FailSafe code to the community.

Tool-augmented language agents frequently fail in real-world deployment due to tool malfunctions--timeouts, API exceptions, or inconsistent outputs--triggering cascading reasoning errors and task abandonment. Existing agent training pipelines optimize only for success trajectories, failing to expose models to the tool failures that dominate real-world usage. We propose \textbf{PALADIN}, a generalizable framework for equipping language agents with robust failure recovery capabilities. PALADIN trains on 50,000+ recovery-annotated trajectories constructed via systematic failure injection and expert demonstrations on an enhanced ToolBench dataset. Training uses LoRA-based fine-tuning to retain base capabilities while injecting recovery competence. At inference, PALADIN detects execution-time errors and retrieves the most similar case from a curated bank of 55+ failure exemplars aligned with ToolScan's taxonomy, then executes the corresponding recovery action. This approach generalizes to novel failures beyond the training distribution, retaining 95.2\% recovery performance on unseen tool APIs. Evaluation across PaladinEval and ToolReflectEval demonstrates consistent improvements in Recovery Rate (RR), Task Success Rate (TSR), Catastrophic Success Rate (CSR), and Efficiency Score (ES). PALADIN improves RR from 32.76% to 89.68% (+57% relative) over ToolBench and outperforms the strongest baseline CRITIC (76.34%) by +13.3%. Against vanilla agents, PALADIN achieves 89.86\% RR (+66% relative improvement from 23.75%). These results establish PALADIN as an effective method for building fault-tolerant agents capable of robust recovery in real-world tool environments.

[Context] Multi-agent reinforcement learning (MARL) has achieved notable success in environments where agents must learn coordinated behaviors. However, transferring knowledge across agents remains challenging in non-stationary environments with changing goals. [Problem] Traditional knowledge transfer methods in MARL struggle to generalize, and agents often require costly retraining to adapt. [Approach] This paper introduces a causal knowledge transfer framework that enables RL agents to learn and share compact causal representations of paths within a non-stationary environment. As the environment changes (new obstacles), agents' collisions require adaptive recovery strategies. We model each collision as a causal intervention instantiated as a sequence of recovery actions (a macro) whose effect corresponds to a causal knowledge of how to circumvent the obstacle while increasing the chances of achieving the agent's goal (maximizing cumulative reward). This recovery action macro is transferred online from a second agent and is applied in a zero-shot fashion, i.e., without retraining, just by querying a lookup model with local context information (collisions). [Results] Our findings reveal two key insights: (1) agents with heterogeneous goals were able to bridge about half of the gap between random exploration and a fully retrained policy when adapting to new environments, and (2) the impact of causal knowledge transfer depends on the interplay between environment complexity and agents' heterogeneous goals.

Among approaches for provably safe reinforcement learning, Model Predictive Shielding (MPS) has proven effective at complex tasks in continuous, high-dimensional state spaces, by leveraging a backup policy to ensure safety when the learned policy attempts to take risky actions. However, while MPS can ensure safety both during and after training, it often hinders task progress due to the conservative and task-oblivious nature of backup policies.This paper introduces Dynamic Model Predictive Shielding (DMPS), which optimizes reinforcement learning objectives while maintaining provable safety. DMPS employs a local planner to dynamically select safe recovery actions that maximize both short-term progress as well as long-term rewards. Crucially, the planner and the neural policy play a synergistic role in DMPS. When planning recovery actions for ensuring safety, the planner utilizes the neural policy to estimate long-term rewards, allowing it to observe beyond its short-term planning horizon.

Traditional anomaly detection techniques onboard satellites are based on reliable, yet limited, thresholding mechanisms which are designed to monitor univariate signals and trigger recovery actions according to specific European Cooperation for Space Standardization (ECSS) standards. However, Artificial Intelligence-based Fault Detection, Isolation and Recovery (FDIR) solutions have recently raised with the prospect to overcome the limitations of these standard methods, expanding the range of detectable failures and improving response times. This paper presents a novel approach to detecting stuck values within the Accelerometer and Inertial Measurement Unit of a drone-like spacecraft for the exploration of Small Solar System Bodies (SSSB), leveraging a multi-channel Convolutional Neural Network (CNN) to perform multi-target classification and independently detect faults in the sensors. Significant attention has been dedicated to ensuring the compatibility of the algorithm within the onboard FDIR system, representing a step forward to the in-orbit validation of a technology that remains experimental until its robustness is thoroughly proven. An integration methodology is proposed to enable the network to effectively detect anomalies and trigger recovery actions at the system level. The detection performances and the capability of the algorithm in reaction triggering are evaluated employing a set of custom-defined detection and system metrics, showing the outstanding performances of the algorithm in performing its FDIR task.

Safety controllers is widely used to achieve safe reinforcement learning. Most methods that apply a safety controller are using handcrafted safety constraints to construct the safety controller. However, when the environment dynamics are sophisticated, handcrafted safety constraints become unavailable. Therefore, it worth to research on constructing safety controllers by learning algorithms. We propose a three-stage architecture for safe reinforcement learning, namely TU-Recovery Architecture. A safety critic and a recovery policy is learned before task training. They form a safety controller to ensure safety in task training. Then a phenomenon induced by disagreement between task policy and recovery policy, called adversarial phenomenon, which reduces learning efficiency and model performance, is described. Auxiliary reward is proposed to mitigate adversarial phenomenon, while help the task policy to learn to recover from high-risk states. A series of experiments are conducted in a robot navigation environment. Experiments demonstrate that TU-Recovery outperforms unconstrained counterpart in both reward gaining and constraint violations during task training, and auxiliary reward further improve TU-Recovery in reward-to-cost ratio by significantly reduce constraint violations.

Experts across multiple states and regions are working on cutting edge science like'call recognition' software to help the shy and elusive Eastern Bristlebird recover the devastating Black Summer bushfires. The state-of-the-art deep learning AI pattern recognition tool is one of eight new recovery projects that have received funding through the Morrison Government's multiregional species coordinator. Minister for the Environment Sussan Ley said the projects will cover a range of recovery actions including the use of scientific surveys to record sightings of the birds to improve understanding of subpopulations and habitat connectivity. "Eastern Bristlebirds are a very secretive bird but can be easily recognised by their melodic song and alarm-call, which is why we are developing new listening tools to support the identification and recovery of this endangered species," Minister Ley said. "By creating a tool that automatically and accurately detects the bird's calls from remote field recordings, and updating radio-transmitter attachment methods, we will be able to track remaining and translocated populations to support their recovery in the future. "We will also be using highly-skilled volunteer scientists to collect data that will inform the future recovery actions we need to take to support the recovery of the Bristlebird across its entire range." Other projects for the Eastern Bristlebird will focus on enhancing recovery through habitat restoration, health and disease research, and support for the establishment of a new genetically viable population in Victoria as a safety net in case of extreme weather events or the spread of disease. "One of the key learnings from the Black Summer bushfires was a need for coordinated on-ground action, monitoring and research, across the entire range of a species, to support its recovery," Minister Ley added. "That is why the Australian Government's $200 million investment in bushfire recovery for wildlife and their habitats is seeing states, territories and stakeholders continuing to work together to support the recovery of ecosystems over a year on from the devastating bushfires.

One of the key challenges to deep reinforcement learning (deep RL) is to ensure safety at both training and testing phases. In this work, we propose a novel technique of unsupervised action planning to improve the safety of on-policy reinforcement learning algorithms, such as trust region policy optimization (TRPO) or proximal policy optimization (PPO). We design our safety-aware reinforcement learning by storing all the history of "recovery" actions that rescue the agent from dangerous situations into a separate "safety" buffer and finding the best recovery action when the agent encounters similar states. Because this functionality requires the algorithm to query similar states, we implement the proposed safety mechanism using an unsupervised learning algorithm, k-means clustering. We evaluate the proposed algorithm on six robotic control tasks that cover navigation and manipulation. Our results show that the proposed safety RL algorithm can achieve higher rewards compared with multiple baselines in both discrete and continuous control problems. The supplemental video can be found at: https://youtu.be/AFTeWSohILo.

A high-precision manipulation task, such as needle threading, is challenging. Physiological studies have proposed connecting low-resolution peripheral vision and fast movement to transport the hand into the vicinity of an object, and using high-resolution foveated vision to achieve the accurate homing of the hand to the object. The results of this study demonstrate that a deep imitation learning based method, inspired by the gaze-based dual resolution visuomotor control system in humans, can solve the needle threading task. First, we recorded the gaze movements of a human operator who was teleoperating a robot. Then, we used only a high-resolution image around the gaze to precisely control the thread position when it was close to the target. We used a low-resolution peripheral image to reach the vicinity of the target. The experimental results obtained in this study demonstrate that the proposed method enables precise manipulation tasks using a general-purpose robot manipulator and improves computational efficiency.