Reinforcement learning with verifiable rewards (RLVR) has shown promise in enhancing the reasoning capabilities of large language models by learning directly from outcome-based rewards. Recent RLVR works that operate under the zero setting avoid supervision in labeling the reasoning process, but still depend on manually curated collections of questions and answers for training. The scarcity of high-quality, human-produced examples raises concerns about the long-term scalability of relying on human supervision, a challenge already evident in the domain of language model pretraining. Furthermore, in a hypothetical future where AI surpasses human intelligence, tasks provided by humans may offer limited learning potential for a superintelligent system. To address these concerns, we propose a new RLVR paradigm called Absolute Zero, in which a single model learns to propose tasks that maximize its own learning progress and improves reasoning by solving them, without relying on any external data. Under this paradigm, we introduce the Absolute Zero Reasoner (AZR), a system that self-evolves its training curriculum and reasoning ability by using a code executor to both validate proposed code reasoning tasks and verify answers, serving as an unified source of verifiable reward to guide open-ended yet grounded learning. Despite being trained entirely without external data, AZR achieves overall SOTA performance on coding and mathematical reasoning tasks, outperforming existing zero-setting models that rely on tens of thousands of in-domain human-curated examples. Furthermore, we demonstrate that AZR can be effectively applied across different model scales and is compatible with various model classes.

Deployment of efficient and accurate Deep Learning models has long been a challenge in autonomous navigation, particularly for real-time applications on resource-constrained edge devices. Edge devices are limited in computing power and memory, making model efficiency and compression essential. In this work, we propose EdgeNavMamba, a reinforcement learning-based framework for goal-directed navigation using an efficient Mamba object detection model. To train and evaluate the detector, we introduce a custom shape detection dataset collected in diverse indoor settings, reflecting visual cues common in real-world navigation. The object detector serves as a pre-processing module, extracting bounding boxes (BBOX) from visual input, which are then passed to an RL policy to control goal-oriented navigation. Experimental results show that the student model achieved a reduction of 67% in size, and up to 73% in energy per inference on edge devices of NVIDIA Jetson Orin Nano and Raspberry Pi 5, while keeping the same performance as the teacher model. EdgeNavMamba also maintains high detection accuracy in MiniWorld and IsaacLab simulators while reducing parameters by 31% compared to the baseline. In the MiniWorld simulator, the navigation policy achieves over 90% success across environments of varying complexity.

The Enterprise Intelligence Platform must integrate logs from numerous third-party vendors in order to perform various downstream tasks. However, vendor documentation is often unavailable at test time. It is either misplaced, mismatched, poorly formatted, or incomplete, which makes schema mapping challenging. We introduce a reinforcement learning agent that can self-improve without labeled examples or model weight updates. During inference, the agent: 1) Identifies ambiguous field-mapping attempts. 2) Generates targeted web-search queries to gather external evidence. 3) Applies a confidence-based reward to iteratively refine its mappings. To demonstrate this concept, we converted Microsoft Defender for Endpoint logs into a common schema. Our method increased mapping accuracy from 56.4\%(LLM-only) to 72.73\%(RAG) to 93.94\% over 100 iterations using GPT-4o. At the same time, it reduced the number of low-confidence mappings requiring expert review by 85\%. This new approach provides an evidence-driven, transparent method for solving future industry problems, paving the way for more robust, accountable, scalable, efficient, flexible, adaptable, and collaborative solutions.

In high-stakes AI applications, even a single action can cause irreparable damage. However, nearly all of sequential decision-making theory assumes that all errors are recoverable (e.g., by bounding rewards). Standard bandit algorithms that explore aggressively may cause irreparable damage when this assumption fails. Some prior work avoids irreparable errors by asking for help from a mentor, but a mentor may not always be available. In this work, we formalize a model of learning with unbounded rewards without a mentor as a two-action contextual bandit with an abstain option: at each round the agent observes an input and chooses either to abstain (always 0 reward) or to commit (execute a preexisting task policy). Committing yields rewards that are upper-bounded but can be arbitrarily negative, and the commit reward is assumed Lipschitz in the input. We propose a caution-based algorithm that learns when not to learn: it chooses a trusted region and commits only where the available evidence does not already certify harm. Under these conditions and i.i.d. inputs, we establish sublinear regret guarantees, theoretically demonstrating the effectiveness of cautious exploration for deploying learning agents safely in high-stakes environments.

Autonomous drone racing (ADR) systems have recently achieved champion-level performance, yet remain highly specific to drone racing. While end-to-end vision-based methods promise broader applicability, no system to date simultaneously achieves full sim-to-real transfer, onboard execution, and champion-level performance. In this work, we present SkyDreamer, to the best of our knowledge, the first end-to-end vision-based ADR policy that maps directly from pixel-level representations to motor commands. SkyDreamer builds on informed Dreamer, a model-based reinforcement learning approach where the world model decodes to privileged information only available during training. By extending this concept to end-to-end vision-based ADR, the world model effectively functions as an implicit state and parameter estimator, greatly improving interpretability. SkyDreamer runs fully onboard without external aid, resolves visual ambiguities by tracking progress using the state decoded from the world model's hidden state, and requires no extrinsic camera calibration, enabling rapid deployment across different drones without retraining. Real-world experiments show that SkyDreamer achieves robust, high-speed flight, executing tight maneuvers such as an inverted loop, a split-S and a ladder, reaching speeds of up to 21 m/s and accelerations of up to 6 g. It further demonstrates a non-trivial visual sim-to-real transfer by operating on poor-quality segmentation masks, and exhibits robustness to battery depletion by accurately estimating the maximum attainable motor RPM and adjusting its flight path in real-time. These results highlight SkyDreamer's adaptability to important aspects of the reality gap, bringing robustness while still achieving extremely high-speed, agile flight.

Testing deep reinforcement learning (DRL) agents in safety-critical domains requires discovering diverse failure scenarios. Existing tools such as INDAGO rely on single-objective optimization focused solely on maximizing failure counts, but this does not ensure discovered scenarios are diverse or reveal distinct error types. We introduce INDAGO-Nexus, a multi-objective search approach that jointly optimizes for failure likelihood and test scenario diversity using multi-objective evolutionary algorithms with multiple diversity metrics and Pareto front selection strategies. We evaluated INDAGO-Nexus on three DRL agents: humanoid walker, self-driving car, and parking agent. On average, INDAGO-Nexus discovers up to 83% and 40% more unique failures (test effectiveness) than INDAGO in the SDC and Parking scenarios, respectively, while reducing time-to-failure by up to 67% across all agents.

In blockchain networks, the strategic ordering of transactions within blocks has emerged as a significant source of profit extraction, known as Maximal Extractable Value (MEV). The transition from spam-based Priority Gas Auctions to structured auction mechanisms like Polygon Atlas has transformed MEV extraction from public bidding wars into sealed-bid competitions under extreme time constraints. While this shift reduces network congestion, it introduces complex strategic challenges where searchers must make optimal bidding decisions within a sub-second window without knowledge of competitor behavior or presence. Traditional game-theoretic approaches struggle in this high-frequency, partially observable environment due to their reliance on complete information and static equilibrium assumptions. We present a reinforcement learning framework for MEV extraction on Polygon Atlas and make three contributions: (1) A novel simulation environment that accurately models the stochastic arrival of arbitrage opportunities and probabilistic competition in Atlas auctions; (2) A PPO-based bidding agent optimized for real-time constraints, capable of adaptive strategy formulation in continuous action spaces while maintaining production-ready inference speeds; (3) Empirical validation demonstrating our history-conditioned agent captures 49\% of available profits when deployed alongside existing searchers and 81\% when replacing the market leader, significantly outperforming static bidding strategies. Our work establishes that reinforcement learning provides a critical advantage in high-frequency MEV environments where traditional optimization methods fail, offering immediate value for industrial participants and protocol designers alike.

Recently, Agentic Reinforcement Learning (Agentic RL) has made significant progress in incentivizing the multi-turn, long-horizon tool-use capabilities of web agents. While mainstream agentic RL algorithms autonomously explore high-uncertainty tool-call steps under the guidance of entropy, excessive reliance on entropy signals can impose further constraints, leading to the training collapse. In this paper, we delve into the challenges caused by entropy and propose the Agentic Entropy-Balanced Policy Optimization (AEPO), an agentic RL algorithm designed to balance entropy in both the rollout and policy update phases. AEPO comprises two core components: (1) a dynamic entropy-balanced rollout mechanism that adaptively allocate global and branch sampling budget through entropy pre-monitoring, while imposing a branch penalty on consecutive high-entropy tool-call steps to prevent over-branching issues; and (2) Entropy-Balanced Policy Optimization that inserts a stop-gradient operation into the high-entropy clipping term to preserve and properly rescale gradients on high-entropy tokens, while incorporating entropy-aware advantage estimation to prioritize learning on high-uncertainty tokens. Results across 14 challenging datasets show that AEPO consistently outperforms 7 mainstream RL algorithms. With just 1K RL samples, Qwen3-14B with AEPO achieves impressive results: 47.6% on GAIA, 11.2% on Humanity's Last Exam, and 43.0% on WebWalker for Pass@1; 65.0% on GAIA, 26.0% on Humanity's Last Exam, and 70.0% on WebWalker for Pass@5. Further analysis reveals that AEPO improves rollout sampling diversity while maintaining stable policy entropy, facilitating scalable web agent training.

This paper proposes a reversible learning framework to improve the robustness and efficiency of value based Reinforcement Learning agents, addressing vulnerability to value overestimation and instability in partially irreversible environments. The framework has two complementary core mechanisms: an empirically derived transition reversibility measure called Phi of s and a, and a selective state rollback operation. We introduce an online per state action estimator called Phi that quantifies the likelihood of returning to a prior state within a fixed horizon K. This measure is used to adjust the penalty term during temporal difference updates dynamically, integrating reversibility awareness directly into the value function. The system also includes a selective rollback operator. When an action yields an expected return markedly lower than its instantaneous estimated value and violates a predefined threshold, the agent is penalized and returns to the preceding state rather than progressing. This interrupts sub optimal high risk trajectories and avoids catastrophic steps. By combining reversibility aware evaluation with targeted rollback, the method improves safety, performance, and stability. In the CliffWalking v0 domain, the framework reduced catastrophic falls by over 99.8 percent and yielded a 55 percent increase in mean episode return. In the Taxi v3 domain, it suppressed illegal actions by greater than or equal to 99.9 percent and achieved a 65.7 percent improvement in cumulative reward, while also sharply reducing reward variance in both environments. Ablation studies confirm that the rollback mechanism is the critical component underlying these safety and performance gains, marking a robust step toward safe and reliable sequential decision making.

Abstract--Imitation learning (IL) aims to learn a policy from expert demonstrations and has been applied to various applications. By learning from the expert policy, IL methods do not require environmental interactions or reward signals. However, most existing imitation learning algorithms assume perfect expert demonstrations, but expert demonstrations often contain imperfections caused by errors from human experts or sensor/control system inaccuracies. T o address the above problems, this work proposes a filter-and-restore framework to best leverage expert demonstrations with inherent noise. Our proposed method first filters clean samples from the demonstrations and then learns conditional diffusion models to recover the noisy ones. We evaluate our proposed framework and existing methods in various domains, including robot arm manipulation, dexterous manipulation, and locomotion. The experiment results show that our proposed framework consistently outperforms existing methods across all the tasks. Ablation studies further validate the effectiveness of each component and demonstrate the framework's robustness to different noise types and levels. These results confirm the practical applicability of our framework to noisy offline demonstration data. MIT A TION learning [1]-[13] aims to learn a policy from expert demonstrations and has been applied to various applications, including robotics [8], industrial automation, strategy board games, video games, etc [14]-[19]. Compared to reinforcement learning (RL), acquiring a policy in a trial-and-error manner, which can be unsafe or expensive, imitation learning (IL) algorithms can learn without environmental interactions. Furthermore, while designing sophisticated RL reward functions is often difficult and tedious [20], [21], IL methods learn from expert demonstrations and do not require reward signals. Despite the wide applicability, most existing imitation learning algorithms assume perfect (i.e., optimal and clean) expert demonstrations, which can be challenging and expensive to collect. Specifically, expert demonstrations often contain imperfections caused by errors from human experts or sensor and control system inaccuracies.