Yihang Yao, Guangtao Zeng, Raina Wu, Yang Zhang, Ding Zhao, Zhang-Wei Hong, Chuang Gan. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Simulation plays a key role in scaling robot learning and validating policies, but constructing simulations remains labor-intensive. In this paper, we introduce ReGen, a generative simulation framework that automates this process using inverse design. Given an agent's behavior (such as a motion trajectory or objective function) and its textual description, we infer the underlying scenarios and environments that could have caused the behavior. Our approach leverages large language models to construct and expand a graph that captures cause-and-effect relationships and relevant entities with properties in the environment, which is then processed to configure a robot simulation environment. Our approach supports (i) augmenting simulations based on ego-agent behaviors, (ii) controllable, counterfactual scenario generation, (iii) reasoning about agent cognition and mental states, and (iv) reasoning with distinct sensing modalities, such as braking due to faulty GPS signals. We demonstrate our method in autonomous driving and robot manipulation tasks, generating more diverse, complex simulated environments compared to existing simulations with high success rates, and enabling controllable generation for corner cases. This approach enhances the validation of robot policies and supports data or simulation augmentation, advancing scalable robot learning for improved generalization and robustness.

Reward shaping is critical in reinforcement learning (RL), particularly for complex tasks where sparse rewards can hinder learning. However, choosing effective shaping rewards from a set of reward functions in a computationally efficient manner remains an open challenge. We propose Online Reward Selection and Policy Optimization (ORSO), a novel approach that frames the selection of shaping reward function as an online model selection problem. ORSO automatically identifies performant shaping reward functions without human intervention with provable regret guarantees. We demonstrate ORSO's effectiveness across various continuous control tasks. Compared to prior approaches, ORSO significantly reduces the amount of data required to evaluate a shaping reward function, resulting in superior data efficiency and a significant reduction in computational time (up to 8×). ORSO consistently identifies high-quality reward functions outperforming prior methods by more than 50% and on average identifies policies as performant as the ones learned using manually engineered reward functions by domain experts.

Large language models (LLMs) have demonstrated remarkable reasoning capabilities across diverse domains. Recent studies have shown that increasing test-time computation enhances LLMs' reasoning capabilities. This typically involves extensive sampling at inference time guided by an external LLM verifier, resulting in a two-player system. Despite external guidance, the effectiveness of this system demonstrates the potential of a single LLM to tackle complex tasks. Thus, we pose a new research problem: *Can we internalize the searching capabilities to fundamentally enhance the reasoning abilities of a single LLM?* This work explores an orthogonal direction focusing on post-training LLMs for autoregressive searching (*i.e.,* an extended reasoning process with self-reflection and self-exploration of new strategies). To achieve this, we propose the Chain-of-Action-Thought (COAT) reasoning and a two-stage training paradigm: 1) a small-scale format tuning stage to internalize the COAT reasoning format and 2) a large-scale self-improvement stage leveraging reinforcement learning. Our approach results in Satori, a 7B LLM trained on open-source models and data. Extensive empirical evaluations demonstrate that Satori achieves state-of-the-art performance on mathematical reasoning benchmarks while exhibits strong generalization to out-of-domain tasks. Code, data, and models are fully open-sourced.

Reinforcement learning (RL) has recently emerged as a compelling approach for enhancing the reasoning capabilities of large language models (LLMs), where an LLM generator serves as a policy guided by a verifier (reward model). However, current RL post-training methods for LLMs typically use verifiers that are fixed (rule-based or frozen pretrained) or trained discriminatively via supervised fine-tuning (SFT). Such designs are susceptible to reward hacking and generalize poorly beyond their training distributions. To overcome these limitations, we propose Tango, a novel framework that uses RL to concurrently train both an LLM generator and a verifier in an interleaved manner. A central innovation of Tango is its generative, process-level LLM verifier, which is trained via RL and co-evolves with the generator. Importantly, the verifier is trained solely based on outcome-level verification correctness rewards without requiring explicit process-level annotations. This generative RL-trained verifier exhibits improved robustness and superior generalization compared to deterministic or SFT-trained verifiers, fostering effective mutual reinforcement with the generator. Extensive experiments demonstrate that both components of Tango achieve state-of-the-art results among 7B/8B-scale models: the generator attains best-in-class performance across five competition-level math benchmarks and four challenging out-of-domain reasoning tasks, while the verifier leads on the ProcessBench dataset. Remarkably, both components exhibit particularly substantial improvements on the most difficult mathematical reasoning problems.

Large language models (LLMs) hold great potential for many natural language applications but risk generating incorrect or toxic content. To probe when an LLM generates unwanted content, the current paradigm is to recruit a $\textit{red team}$ of human testers to design input prompts (i.e., test cases) that elicit undesirable responses from LLMs. However, relying solely on human testers is expensive and time-consuming. Recent works automate red teaming by training a separate red team LLM with reinforcement learning (RL) to generate test cases that maximize the chance of eliciting undesirable responses from the target LLM. However, current RL methods are only able to generate a small number of effective test cases resulting in a low coverage of the span of prompts that elicit undesirable responses from the target LLM. To overcome this limitation, we draw a connection between the problem of increasing the coverage of generated test cases and the well-studied approach of curiosity-driven exploration that optimizes for novelty. Our method of curiosity-driven red teaming (CRT) achieves greater coverage of test cases while mantaining or increasing their effectiveness compared to existing methods. Our method, CRT successfully provokes toxic responses from LLaMA2 model that has been heavily fine-tuned using human preferences to avoid toxic outputs. Code is available at https://github.com/Improbable-AI/curiosity_redteam.

The ability to efficiently explore high-dimensional state spaces is essential for the practical success of deep Reinforcement Learning (RL). This paper introduces a new exploration technique called Random Latent Exploration (RLE), that combines the strengths of exploration bonuses and randomized value functions (two popular approaches for effective exploration in deep RL). RLE leverages the idea of perturbing rewards by adding structured random rewards to the original task rewards in certain (random) states of the environment, to encourage the agent to explore the environment during training. RLE is straightforward to implement and performs well in practice. To demonstrate the practical effectiveness of RLE, we evaluate it on the challenging Atari and IsaacGym benchmarks and show that RLE exhibits higher overall scores across all the tasks than other approaches, including action-noise and randomized value function exploration.

Reinforcement Learning (RL) has been a powerful tool for training robots to acquire agile locomotion skills. To learn locomotion, it is commonly necessary to introduce additional reward-shaping terms, such as an energy minimization term, to guide an algorithm like Proximal Policy Optimization (PPO) to good performance. Prior works rely on hyper-parameter tuning on the weight of the reward shaping terms to obtain satisfactory task performance. To save the efforts of tuning these weights, we adopt the Extrinsic-Intrinsic Policy Optimization (EIPO) framework. The key idea of EIPO is to establish a constrained optimization framework for the primary objective of enhancing task performance and the secondary objective of minimizing energy consumption. It seeks a policy that minimizes the energy consumption objective within the optimal policy space for task performance. This guarantees that the learned policy excels in task performance while conserving energy, all without requiring manual weight adjustments for both objectives. Our experiments evaluate EIPO on various quadruped locomotion tasks, revealing that policies trained with EIPO consistently achieve higher task performance than PPO comparisons while maintaining comparable energy consumption levels. Furthermore, EIPO exhibits superior task performance in real-world evaluations compared to PPO.

Generating varied scenarios through simulation is crucial for training and evaluating safety-critical systems, such as autonomous vehicles. Yet, the task of modeling the trajectories of other vehicles to simulate diverse and meaningful close interactions remains prohibitively costly. Adopting language descriptions to generate driving behaviors emerges as a promising strategy, offering a scalable and intuitive method for human operators to simulate a wide range of driving interactions. However, the scarcity of large-scale annotated language-trajectory data makes this approach challenging. To address this gap, we propose Text-to-Drive (T2D) to synthesize diverse driving behaviors via Large Language Models (LLMs). We introduce a knowledge-driven approach that operates in two stages. In the first stage, we employ the embedded knowledge of LLMs to generate diverse language descriptions of driving behaviors for a scene. Then, we leverage LLM’s reasoning capabilities to synthesize these behaviors in simulation. At its core, T2D employs an LLM to construct a state chart that maps low-level states to high-level abstractions. This strategy aids in downstream tasks such as summarizing low-level observations, assessing policy alignment with behavior description, and shaping the auxiliary reward, all without needing human supervision. With our knowledge-driven approach, we demonstrate that T2D generates more diverse trajectories compared to other baselines and offers a natural language interface that allows for interactive incorporation of human preference. Please check our website for more examples: here

In many reinforcement learning (RL) applications, incorporating heuristic rewards alongside the task reward is crucial for achieving desirable performance. Heuristics encode prior human knowledge about how a task should be done, providing valuable hints for RL algorithms. However, such hints may not be optimal, limiting the performance of learned policies. The currently established way of using heuristics is to modify the heuristic reward in a manner that ensures that the optimal policy learned with it remains the same as the optimal policy for the task reward (i.e., optimal policy invariance). However, these methods often fail in practical scenarios with limited training data. We found that while optimal policy invariance ensures convergence to the best policy based on task rewards, it doesn't guarantee better performance than policies trained with biased heuristics under a finite data regime, which is impractical. In this paper, we introduce a new principle tailored for finite data settings. Instead of enforcing optimal policy invariance, we train a policy that combines task and heuristic rewards and ensures it outperforms the heuristic-trained policy. As such, we prevent policies from merely exploiting heuristic rewards without improving the task reward. Our experiments on robotic locomotion, helicopter control, and manipulation tasks demonstrate that our method consistently outperforms the heuristic policy, regardless of the heuristic rewards' quality. Code is available at https://github.com/Improbable-AI/hepo.

Macro actions have been demonstrated to be beneficial for the learning processes of an agent and have encouraged a variety of techniques to be developed for constructing more effective ones. However, previous techniques usually do not further consider combining macro actions to form a synergistic macro action ensemble, in which synergism exhibits when the constituent macro actions are favorable to be jointly used by an agent during evaluation. Such a synergistic macro action ensemble may potentially allow an agent to perform even better than the individual macro actions within it. Motivated by the recent advances of neural architecture search (NAS), in this brief, we formulate the construction of a synergistic macro action ensemble as a Markov decision process (MDP) and evaluate the constructed macro action ensemble as a whole. Such a problem formulation enables synergism to be taken into account by the proposed evaluation procedure. Our experimental results demonstrate that the proposed framework is able to discover the synergistic macro action ensembles. Furthermore, we also highlight the benefits of these macro action ensembles through a set of analytical cases.

Reinforcement learning is time-consuming for complex tasks due to the need for large amounts of training data. Recent advances in GPU-based simulation, such as Isaac Gym, have sped up data collection thousands of times on a commodity GPU. Most prior works have used on-policy methods like PPO due to their simplicity and easy-to-scale nature. Off-policy methods are more sample-efficient, but challenging to scale, resulting in a longer wall-clock training time. This paper presents a novel Parallel Q-Learning (PQL) scheme that outperforms PPO in terms of wall-clock time and maintains superior sample efficiency. The driving force lies in the parallelization of data collection, policy function learning, and value function learning. Different from prior works on distributed off-policy learning, such as Apex, our scheme is designed specifically for massively parallel GPU-based simulation and optimized to work on a single workstation. In experiments, we demonstrate the capability of scaling up Q-learning methods to tens of thousands of parallel environments and investigate important factors that can affect learning speed, including the number of parallel environments, exploration strategies, batch size, GPU models, etc. The code is available at https://github.com/Improbable-AI/pql.

We consider solving sequential decision-making problems in the scenario where the agent has access to two supervision sources: $\textit{reward signal}$ and a $\textit{teacher}$ that can be queried to obtain a $\textit{good}$ action for any state encountered by the agent. Learning solely from rewards, or reinforcement learning, is data inefficient and may not learn high-reward policies in challenging scenarios involving sparse rewards or partial observability. On the other hand, learning from a teacher may sometimes be infeasible. For instance, the actions provided by a teacher with privileged information may be unlearnable by an agent with limited information (i.e., partial observability). In other scenarios, the teacher might be sub-optimal, and imitating their actions can limit the agent’s performance. To overcome these challenges, prior work proposed to jointly optimize imitation and reinforcement learning objectives but relied on heuristics and problem-specific hyper-parameter tuning to balance the two objectives. We introduce Teacher Guided Reinforcement Learning (TGRL), a principled approach to dynamically balance following the teacher’s guidance and leveraging RL. TGRL outperforms strong baselines across diverse domains without hyperparameter tuning.

Offline reinforcement learning (RL) enables learning a decision-making policy without interaction with the environment. This makes it particularly beneficial in situations where such interactions are costly. However, a known challenge for offline RL algorithms is the distributional mismatch between the state-action distributions of the learned policy and the dataset, which can significantly impact performance. State-of-the-art algorithms address it by constraining the policy to align with the state-action pairs in the dataset. However, this strategy struggles on datasets that predominantly consist of trajectories collected by low-performing policies and only a few trajectories from high-performing ones. Indeed, the constraint to align with the data leads the policy to imitate low-performing behaviors predominating the dataset. Our key insight to address this issue is to constrain the policy to the policy that collected the good parts of the dataset rather than all data. To this end, we optimize the importance sampling weights to emulate sampling data from a data distribution generated by a nearly optimal policy. Our method exhibits considerable performance gains (up to five times better) over the existing approaches in state-of-the-art offline RL algorithms over 72 imbalanced datasets with varying types of imbalance.

We present a strong baseline that surpasses the performance of previously published methods on the Habitat Challenge task of navigating to a target object in indoor environments. Our method is motivated from primary failure modes of prior state-of-the-art: poor exploration, inaccurate object identification, and agent getting trapped due to imprecise map construction. We make three contributions to mitigate these issues: (i) First, we show that existing map-based methods fail to effectively use semantic clues for exploration. We present a semantic-agnostic exploration strategy (called Stubborn) without any learning that surprisingly outperforms prior work. (ii) We propose a strategy for integrating temporal information to improve object identification. (iii) Lastly, due to inaccurate depth observation the agent often gets trapped in small regions. We develop a multi-scale collision map for obstacle identification that mitigates this issue. Website: https://github.com/Improbable-AI/Stubborn

State-of-the-art reinforcement learning (RL) algorithms typically use random sampling (e.g., $\epsilon$-greedy) for exploration, but this method fails on hard exploration tasks like Montezuma's Revenge. To address the challenge of exploration, prior works incentivize exploration by rewarding the agent when it visits novel states. Such intrinsic rewards (also called exploration bonus or curiosity) often lead to excellent performance on hard exploration tasks. However, on easy exploration tasks, the agent gets distracted by intrinsic rewards and performs unnecessary exploration even when sufficient task (also called extrinsic) reward is available. Consequently, such an overly curious agent performs worse than an agent trained with only task reward. Such inconsistency in performance across tasks prevents the widespread use of intrinsic rewards with RL algorithms. We propose a principled constrained optimization procedure called Extrinsic-Intrinsic Policy Optimization (EIPO) that automatically tunes the importance of the intrinsic reward: it suppresses the intrinsic reward when exploration is unnecessary and increases it when exploration is required. The results is superior exploration that does not require manual tuning in balancing the intrinsic reward against the task reward. Consistent performance gains across sixty-one ATARI games validate our claim. The code is available at https://github.com/Improbable-AI/eipo.

Deep reinforcement learning (DRL) has been demonstrated to provide promising results in several challenging decision making and control tasks. However, the required inference costs of deep neural networks (DNNs) could prevent DRL from being applied to mobile robots which cannot afford high energy-consuming computations. To enable DRL methods to be affordable in such energy-limited platforms, we propose an asymmetric architecture that reduces the overall inference costs via switching between a computationally expensive policy and an economic one. The experimental results evaluated on a number of representative benchmark suites for robotic control tasks demonstrate that our method is able to reduce the inference costs while retaining the agent’s overall performance.

Collecting training data from the physical world is usually time-consuming and even dangerous for fragile robots, and thus, recent advances in robot learning advocate the use of simulators as the training platform. Unfortunately, the reality gap between synthetic and real visual data prohibits direct migration of the models trained in virtual worlds to the real world. This paper proposes a modular architecture for tackling the virtual-to-real problem. The proposed architecture separates the learning model into a perception module and a control policy module, and uses semantic image segmentation as the meta representation for relating these two modules. The perception module translates the perceived RGB image to semantic image segmentation. The control policy module is implemented as a deep reinforcement learning agent, which performs actions based on the translated image segmentation. Our architecture is evaluated in an obstacle avoidance task and a target following task. Experimental results show that our architecture significantly outperforms all of the baseline methods in both virtual and real environments, and demonstrates a faster learning curve than them. We also present a detailed analysis for a variety of variant configurations, and validate the transferability of our modular architecture.

Efficient exploration remains a challenging research problem in reinforcement learning, especially when an environment contains large state spaces, deceptive local optima, or sparse rewards. To tackle this problem, we present a diversity-driven approach for exploration, which can be easily combined with both off- and on-policy reinforcement learning algorithms. We show that by simply adding a distance measure to the loss function, the proposed methodology significantly enhances an agent's exploratory behaviors, and thus preventing the policy from being trapped in local optima. We further propose an adaptive scaling method for stabilizing the learning process. We demonstrate the effectiveness of our method in huge 2D gridworlds and a variety of benchmark environments, including Atari 2600 and MuJoCo. Experimental results show that our method outperforms baseline approaches in most tasks in terms of mean scores and exploration efficiency.

We introduce two tactics, namely the strategically-timed attack and the enchanting attack, to attack reinforcement learning agents trained by deep reinforcement learning algorithms using adversarial examples. In the strategically-timed attack, the adversary aims at minimizing the agent's reward by only attacking the agent at a small subset of time steps in an episode. Limiting the attack activity to this subset helps prevent detection of the attack by the agent. We propose a novel method to determine when an adversarial example should be crafted and applied. In the enchanting attack, the adversary aims at luring the agent to a designated target state. This is achieved by combining a generative model and a planning algorithm: while the generative model predicts the future states, the planning algorithm generates a preferred sequence of actions for luring the agent. A sequence of adversarial examples is then crafted to lure the agent to take the preferred sequence of actions. We apply the proposed tactics to the agents trained by the state-of-the-art deep reinforcement learning algorithm including DQN and A3C. In 5 Atari games, our strategically-timed attack reduces as much reward as the uniform attack (i.e., attacking at every time step) does by attacking the agent 4 times less often. Our enchanting attack lures the agent toward designated target states with a more than 70% success rate. Example videos are available at http://yclin.me/adversarial_attack_RL/.